Hi everyone.
I am working on a fluid dynamics project processing high-speed images of sprays to calculate the global spray cone angle (SCA). I am using MATLAB to binarize the images, extract the boundaries of the liquid core, and apply a linear regression to the left and right edges to find the angle.
However, I am facing a high number of severe outliers in my angle calculations, such as sudden spikes or drops across frames. From my analysis, these occur due to disconnected droplets or liquid ligaments that trick the edge-detection, pulling the regression line outwards. I need a systematic criterion to dynamically set the lower limit of my region of interest for every single image.
I would like to ask the community how you typically filter or prevent these structural outliers in your experience with sprays. Would mathematically identifying the exact break-up line and strictly limiting the regression to the intact liquid core above it be the definitive solution to stabilize the angle? If so, what is the most robust algorithmic approach in MATLAB to systematically detect this break-up line across varying pressures, without being easily fooled by isolated droplets or contrast errors?
I would deeply appreciate any insights, logic suggestions, or functions you could recommend for this challenge. Thank you in advance!Hi everyone.
I am working on a fluid dynamics project processing high-speed images of sprays to calculate the global spray cone angle (SCA). I am using MATLAB to binarize the images, extract the boundaries of the liquid core, and apply a linear regression to the left and right edges to find the angle.
However, I am facing a high number of severe outliers in my angle calculations, such as sudden spikes or drops across frames. From my analysis, these occur due to disconnected droplets or liquid ligaments that trick the edge-detection, pulling the regression line outwards. I need a systematic criterion to dynamically set the lower limit of my region of interest for every single image.
I would like to ask the community how you typically filter or prevent these structural outliers in your experience with sprays. Would mathematically identifying the exact break-up line and strictly limiting the regression to the intact liquid core above it be the definitive solution to stabilize the angle? If so, what is the most robust algorithmic approach in MATLAB to systematically detect this break-up line across varying pressures, without being easily fooled by isolated droplets or contrast errors?
I would deeply appreciate any insights, logic suggestions, or functions you could recommend for this challenge. Thank you in advance! Hi everyone.
I am working on a fluid dynamics project processing high-speed images of sprays to calculate the global spray cone angle (SCA). I am using MATLAB to binarize the images, extract the boundaries of the liquid core, and apply a linear regression to the left and right edges to find the angle.
However, I am facing a high number of severe outliers in my angle calculations, such as sudden spikes or drops across frames. From my analysis, these occur due to disconnected droplets or liquid ligaments that trick the edge-detection, pulling the regression line outwards. I need a systematic criterion to dynamically set the lower limit of my region of interest for every single image.
I would like to ask the community how you typically filter or prevent these structural outliers in your experience with sprays. Would mathematically identifying the exact break-up line and strictly limiting the regression to the intact liquid core above it be the definitive solution to stabilize the angle? If so, what is the most robust algorithmic approach in MATLAB to systematically detect this break-up line across varying pressures, without being easily fooled by isolated droplets or contrast errors?
I would deeply appreciate any insights, logic suggestions, or functions you could recommend for this challenge. Thank you in advance! sprayangle, help, image processing MATLAB Answers — New Questions
