In the last two decades, the interest towards Digital Image Correlation (DIC) has grown steadily and in various scientific sectors, this is thanks to its reduced cost, ease of implementation and to its contact-less nature. While most of the application concern static measures such as strain analysis and material characterization, recently new employments for DIC are catching up. In particular, the possibility to carry out measurements in presence of relative motion between the camera and the target has been investigated as an extension of the applications field of this technique. With respect to static measures, in dynamic conditions an additional source of uncertainty is present, this is represented by the motion blur effect, which is experienced whenever the relative displacement between the camera and the target during the exposure time is not negligible. Motion blur is a source of uncertainty in that it decreases the contrast in the image and the definition of the speckles in the acquired pattern, making it difficult to identify the subsets in the images of the deformed target during the DIC analysis. For this reason different de-blurring algorithms have been developed. In this work the estimation and compensation algorithm exposed in [1] is considered. At first an attempt of improving the motion blur estimation phase was made basing on the extraction of the main lobe of the Optical Transfer Function (OTF). Afterwards, the compensation phase was analyzed, in this case two different estimations for the Noise to Signal Ratio (NSR) were investigated. Eventually the robustness of the algorithm was tested both though synthetically blurred images and through the acquisition of vibrating targets in quasi-real conditions.
Negli ultimi vent'anni, l'interesse verso la Digital Image Correlation è cresciuto in modo stabile in diversi settori scientifici, ciò è grazie al suo costo ridotto, facilità di implementazione e alla possibilità di condurre misure senza contatto. Mentre la maggior parte delle applicazioni della DIC concerne misure statiche come analisi delle deformazioni e caratterizzazione dei materiali, recentemente stanno emergendo nuovi impieghi per questa tecnica di misura. In particolare, la possibilità di effettuare misure in presenza di moto relativo non trascurabile tra fotocamera e target è stata esplorata come ampliamento del campo di applicazione della DIC. Rispetto alle misure statiche, in condizioni dinamiche è presente una fonte d'incertezza aggiuntiva che è rappresentata dall'effetto mosso, il quale si verifica ogni volta che lo spostamento relativo tra la fotocamera e il target durante il tempo di esposizione non è trascurabile. L'effetto mosso è fonte di incertezza in quanto riduce il contrasto nelle immagini e la definizione degli speckle, rendendo difficoltosa l'identificazione dei subset nelle immagini del target deformato durante l'analisi DIC. Per questo motivo, in tempi recenti vari algoritmi di compensazione sono stati sviluppati. In questo lavoro, l'algoritmo di stima e compensazione sviluppato in [1] è stato preso in considerazione. Inizialmente è stato portato avanti un tentativo di miglioramento nella fase di stima basato sull'estrazione del lobo centrale della Funzione di Trasferimento Ottica (OTF). Dopodiché la fase di compensazione è stata analizzata, a questo proposito due stime diverse per il rapporto rumore segnale (NSR) sono state esplorate. Infine la robustezza dell'algoritmo considerato è stata messa alla prova sia attraverso l'uso di immagini mosse sinteticamente, sia attraverso l'acquisizione di immagini di target vibranti in condizioni simili alle reali.
Metrological characterization of a motion blur estimation and compensation algorithm for dynamic digital image correlation measurements
Sala, Giovanni
2020/2021
Abstract
In the last two decades, the interest towards Digital Image Correlation (DIC) has grown steadily and in various scientific sectors, this is thanks to its reduced cost, ease of implementation and to its contact-less nature. While most of the application concern static measures such as strain analysis and material characterization, recently new employments for DIC are catching up. In particular, the possibility to carry out measurements in presence of relative motion between the camera and the target has been investigated as an extension of the applications field of this technique. With respect to static measures, in dynamic conditions an additional source of uncertainty is present, this is represented by the motion blur effect, which is experienced whenever the relative displacement between the camera and the target during the exposure time is not negligible. Motion blur is a source of uncertainty in that it decreases the contrast in the image and the definition of the speckles in the acquired pattern, making it difficult to identify the subsets in the images of the deformed target during the DIC analysis. For this reason different de-blurring algorithms have been developed. In this work the estimation and compensation algorithm exposed in [1] is considered. At first an attempt of improving the motion blur estimation phase was made basing on the extraction of the main lobe of the Optical Transfer Function (OTF). Afterwards, the compensation phase was analyzed, in this case two different estimations for the Noise to Signal Ratio (NSR) were investigated. Eventually the robustness of the algorithm was tested both though synthetically blurred images and through the acquisition of vibrating targets in quasi-real conditions.File | Dimensione | Formato | |
---|---|---|---|
2021_7_Sala.pdf
accessibile in internet per tutti
Descrizione: Testo della tesi aggiornato
Dimensione
17.97 MB
Formato
Adobe PDF
|
17.97 MB | Adobe PDF | Visualizza/Apri |
I documenti in POLITesi sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/10589/177973