Using Image Processing in MATLAB to Analyze Cavitation of a Micro Hydo-Turbine
Mentor 1
Dr. Amano
Location
Union 250
Start Date
5-4-2019 12:00 PM
Description
Image processing is a powerful tool for effective data analysis. The particular areas of interest for image processing in this experiment were to find the total area of cavitation around the turbine blade from multiple experiments and Images taken from a Computational Fluid Dynamics(CFD) program as well as finding a way to display the cavitation in a more visible manner without compromising the integrity of the image. There are two particular areas of interest for image processing in this experiment: 1) Find the total area of cavitation around the turbine blade from multiple experiments and images taken from a Computational Fluid Dynamics (CFD) program; 2) Find a way to display the cavitation in a more visible manner without compromising the integrity of the image. Issues encountered when searching for the area of cavitation from experimental data were the following: 1) Movement of the camera to a different location each time data was collected; 2) Finding the appropriate threshold value for converting the images into binary images without ruining the integrity of the data represented within each image. To improve upon the re-usability of the script, object recognition was investigated as a next step as a method for “auto-scaling” the images from different experimental test runs. For the image processing, respectively to 1 and 2 areas of interest; 1) The CFD images were cropped around an area of interest then the turbine was subtracted out of the image, leaving behind just the cavitation. The percentage of cavitation about the turbine was then found by taking the area of cavitation divided by the total area of the turbine blade. 2) To make the experimental images more visible, a colormap was applied to the image. The lighter regions of the image are given lighter colors while the darker regions are given darker colors.
Using Image Processing in MATLAB to Analyze Cavitation of a Micro Hydo-Turbine
Union 250
Image processing is a powerful tool for effective data analysis. The particular areas of interest for image processing in this experiment were to find the total area of cavitation around the turbine blade from multiple experiments and Images taken from a Computational Fluid Dynamics(CFD) program as well as finding a way to display the cavitation in a more visible manner without compromising the integrity of the image. There are two particular areas of interest for image processing in this experiment: 1) Find the total area of cavitation around the turbine blade from multiple experiments and images taken from a Computational Fluid Dynamics (CFD) program; 2) Find a way to display the cavitation in a more visible manner without compromising the integrity of the image. Issues encountered when searching for the area of cavitation from experimental data were the following: 1) Movement of the camera to a different location each time data was collected; 2) Finding the appropriate threshold value for converting the images into binary images without ruining the integrity of the data represented within each image. To improve upon the re-usability of the script, object recognition was investigated as a next step as a method for “auto-scaling” the images from different experimental test runs. For the image processing, respectively to 1 and 2 areas of interest; 1) The CFD images were cropped around an area of interest then the turbine was subtracted out of the image, leaving behind just the cavitation. The percentage of cavitation about the turbine was then found by taking the area of cavitation divided by the total area of the turbine blade. 2) To make the experimental images more visible, a colormap was applied to the image. The lighter regions of the image are given lighter colors while the darker regions are given darker colors.
