Activity for JPIV
-
Peter Vennemann posted a comment on discussion JPIV Open Discussion
Dear Joseph, I just implemented background substraction into JPIV directly. With the actual version, you don't need ImageJ any more. https://eguvep.github.io/jpiv/ Regards, Peter
-
Dear JPIV users, Source code and documentation of JPIV moved to GitHub. Please visit the updated project at: https://eguvep.github.io/jpiv/index.html There are two main reasons for moving to GitHub: I updated and moved the documentation from my private homepage (jpiv.vennemann-online.de) to GitHub Pages (https://eguvep.github.io/jpiv/index.html), to ease maintenance with git, and to enable contributions by others. JPIV and OpenPiv (https://github.com/OpenPIV) started to collaborate. Forking and merging...
-
Try png as a file format. That is the most stable export option on most platforms.
-
Dear Javi, JPIV is just for 2D flow fields, not for Stereo PIV or Tomo-PIV. The reconstruction of 3D flow fields from Multi-Plane 2D PIV (Scanning PIV) is implemented, though (for incompressible flows). Regards, Peter
-
Dear Megan, It is just the peak hight, not the signal to noise ratio, in the fifth column, although they are kind of propotional. There is no need to export it, because jvc files are just space separated text files. You can open them with any text editor or any other visualization tool. Regards, Peter
-
Trajectory plot is not implemented in JPIV. I would use https://matplotlib.org/3.1.3/api/_as_gen/matplotlib.axes.Axes.streamplot.html#matplotlib.axes.Axes.streamplot.
-
There seems to be an error in the .jvc file. Please review the file with a text editor.
-
Dear Karhikeyan, if you know the mm/px value, you can directly enter it as JPIV - Preferences - Vector Plot - Axis Unit Conversion Factor. You can divide this value by the time difference between the frames and enter it as a Unit Conversion Factor of the color legend, to get real velocity units. In JPIV, the orginal data is never changed. Just the labels of the color bar are changed. Regards, Peter
-
Dear Joseph, the problem is solved. It is neither a bug nor a physical effect. It is a background effect of the images. The correlation does not only indicate particle movement, but also brightness-gradients. In the first pass, the velocity information dominates. In the second pass, the evaluation is dominated by the bright spots in the background (where erythrocytes are absent). I removed the background of the images by a rolling ball algorithm (ImageJ, radius 2 pixel, see zipfile). This alomost...
-
Joseph Sherwood posted a comment on discussion JPIV Open Discussion
Hi Peter, thanks for thinking about this. I can't be sure, but I don't believe that the finitie particles size is the case, given the spatial and temporal averaging. Furthermore, I had previously processed the same data and got results that did not produce this response. According to my notes, I used these settings 64 16 8 8 8 8 64 16 8 8 8 8 4 4 4 4 4 4 with a pre-shift of 16 pixels. I also have both RBC images and quasi-simultaneously acquired (alternate pairs) images of fluorescent microparticles....
-
Still thinking... There are about as many velocity groups (bins) as there are particles over the channel width (inclusive overlap, see attachement). Are there preferred velocities in narrow channels? If the channel is just a little bit wider than one blood cell, there is just one velocity. If the channel is much wider, there are no preferred velocities. But you are inbetween...
-
Dear Joseph, I used the same methods to analyse your vector data as I used to analyse the synthetic image vector data. (See attachements). The blood cell velocity profiles of the two pass evaluation is indeed less smooth than the single pass evaluation result – as you've shown above, already. The profiles look kind of binned, although the histrograms do not show any binning towards integer pixel values. The do show binning towards certain displacements, but this is mainly because of the horizontal...
-
Dear Peter, Sorry for the delay, last term was a bit hectic. I have tried your images and they work fine on my system. I attach a stack of images of blood flow in a microchannel. Maybe this issue down to some sampling issue in the blood, but I'm not clear. It didn't happen to me in the past, so it is strange. I couldn't find the jpiv_settings.ser, so have included a screen shot. Thanks for your help, Kind regards, Joseph
-
Dear Peter, Sorry for the delay, last term was a bit hectic. I have tried you images and they work fine on my system. I attach a stack of images of blood flow in a microchannel. Maybe this is down to some sampling issue in the blood, but I'm not clear. It didn't happen to me in the past, so it is strange. I couldn't find the jpiv_settings.ser, so have included a screen shot. Thanks for your help, Kind regards, Joseph
-
Dear Joseph, I generated some synthetic data. As far as I can see, there is no aliasing or discretization. I just dump everything here, that I tested. Possibly, you can reproduce the effect with your settings on your machine on the synthetic images. Regards, Peter
-
Dear Joseph, That effect gives me a hard time to think about. I am still not sure, about it. I checked my code, but couldn't find a bug, yet. Can you exclude any physical reason, for example the diameter or depth of the channel varying between two values or kind of an oscilating velocity profile? You may post your images and the settings-file (jpiv_settings.ser) here, so that I could further investigate that issue. Regards, Peter
-
Hi Peter, Long time no JPIV. I have recently returned to some old data on blood cells in a microchannel. I have tried reprocessing the data with JPIV and am getting an odd responses. After much investigation, it appears the issue is with multi-pass. Find atatched some jvc files from processing the same data set (60 image pairs, using ensemble averaging, 50% overlap) with 8 or 4 pixel vector spacing using 1 or 2 passes of the same window parameters. In the first pass, the data are quite smooth. After...
-
Tom Nakzy modified a comment on discussion Users and Publications
I just started manipulating images. I tried to obtain the velocity field using JPIV. This question is similar to a post titled "Image processing area" at 2017-08-01. I processed images of 1500 x 1300. Output data jvc file has data in the range (64, 64) to (1344, 1216). Interrogation window is 128x128 at first pass, and at 3pass, width is 16x16. I think this gap is considerably larger than half of the window width. I also analyzed even smaller images with PIV. However, I thought that the gap was large...
-
I just started manipulating images. I tried to obtain the velocity field using JPIV. This question is similar to a post titled "Image processing area" at 2017-08-01. I processed images of 1500 x 1300. Output data jvc file has data in the range (64, 64) to (1344, 1216). Interrogation window is 128x128 at first pass, and at 3pass, width is 16x16. I think this gap is considerably larger than half of the window width. I also analyzed even smaller images with PIV. However, I thought that the gap was large...
-
I just started manipulating images. I tried to obtain the velocity field using JPIV. This question is similar to a post titled "Image processing area" at 2017-08-01. I processed images of 1500 x 1300. Output data jvc file has data in the range (64, 64) to (1344, 1216). Interrogation window is 128x128 at first pass, and at 3pass, width is 16x16. I think this gap is considerably larger than half of the window width. I also analyzed even smaller images with PIV. However, I thought that the gap was large...
-
I just started manipulating images. I tried to obtain the velocity field using JPIV. This question is similar to a post titled "Image processing area" at 2017-08-01. I processed images of 1500 x 1300. Output data jvc file has data in the range (64, 64) to (1344, 1216). Interrogation window is 128x128 at first pass, and at 3pass, width is 16x16. I think this gap is considerably larger than half of the window width. I also analyzed even smaller images with PIV. However, I thought that the gap was equally...
-
I just started manipulating images. I tried to obtain the velocity field using JPIV. This question is similar to a post titled "Image processing area" at 2017-08-01. I processed images of 1500 x 1300. Output data jvc file has data in the range (64, 64) to (1344, 1216). Interrogation window is 128x128 at first pass, and at 3pass, width is 16x16. I think this gap is considerably larger than half of the window width. What is the cause of this situation? Which parameters in the configuration window are...
-
2017-08-01 It is similar to a post titled "Image processing area" in 2017. I processed images of 1500 x 1300. Output data jvc file has data in the range (64, 64) to (1344, 1216). Interrogation window is 128x128, 3pass width is 16x16. I think this is considerably larger than half of the window width. What is the cause of this situation? Which parameters in the configuration window are involved?
-
Dear all, JPIV users who experience out of memory problems may try the actual version 18.03.27. Background: I changed the data type of the correlation map to a BufferedImage of type float. Up to now it had been a PlanarImage of type float. The advantage of the PlanarImage was a slightly lower computation time. The backdraw was higher memory usage. In recent years, CPU power and the size of digital images, both went up. Low memory usage thus became more important than low CPU load. Under actual circumstances...
-
Ben modified a comment on discussion JPIV Open Discussion
Dear Peter, thanks for clarifying! Regards, Ben
-
Dear Peter, thanks for clarifying! Thanks again! Regards, Ben
-
karthikeyan posted a comment on discussion JPIV Open Discussion
In my case I am not aware of the magnification but I know the mm/px value. How to incorporate this into the unit conversion section and also for axis ticking? NOTE: I am aware of my camera px size. Hence my sensor size and the field of view size. These are what is available to me.
-
Dear John, JPIV just calculates the third velocity component for incompressible flows. 3D-visualization is not implemented. Use TECPLOT or PARAVIEW or some other general purpose visualization tool for that (I used TECPLOT). Regards! Peter
-
John Sheppard posted a comment on discussion JPIV Open Discussion
Dear Peter, I read your thesis. But still I can't find how to get 3D PIV image. I did as your answer at June 28th, and I got images. But those images seem same with original ones. How can I have the 3D image like the one shown in your web page? Don't you have a manual or something like that? I wish to know every single click. Regards, Sheppard
-
Dear Peter, I could get the thesis. I'm sorry for many troubles. Thank you. Regards, Sheppard
-
Dear Sheppard, Now I see the problem: For whatever reason, firefox includes the period at the end of the sentence into the URL. Just remove it, and you can access the thesis. Regards, Peter
-
Does it mean I have to access your thesis in a few days since you upload it? Then would you upload it again? I'll be careful this time not to miss. Regards, Sheppard
-
Dear Maria Luisa, Have a look at the interrogation window table: The first two rows specify width and height of the interrogation windows. Powers of two are faster. The third and fourth row specify the search domain. This is an area in the middle of the interrogation window. Choose a value that is little larger than twice the maximum expected discplacement. Peaks outside this area will not be accounted for. Your question is about the last two rows: If the vector spacing is choosen equal to the interrogation...
-
Dear John, It seems, they constantly move things around at the TU Delft library: Here it is: http://resolver.tudelft.nl/uuid:28a0bbb8-e671-4806-ba52-cca2e6db0d50. Regards, Peter
-
Dear Karthikeyan, It woud result in a higer value, but would not produce a peak, just a uniform (or pyramid like) function. A correlation result is »high« if the signal to noise ratio (peak height divided by average heigth) is »high« or at least more than 1.3. Regards, Peter
-
Dear Maria Luisa, Thanks a lot for sharing your interesting poster! Congratulations also for this very impressive work: Aligning and rescaling was tedious work, I guess. And: More than 300 km/h – wow! Do you know the density of the atmosphere in the evaluated cloud layer? You might want to get rid of the exponential format at the color bar. Just modify the format string, as described here: 00 prints exactly two digits, #0.00 prints at least one digit in front of the decimal dot and exactly two digits...
-
Poster by M. L. Moriconi: https://sourceforge.net/p/jpiv/discussion/950409/thread/16f3c947/15a9/attachment/Poster2.jpg
-
Francois modified a comment on discussion JPIV Open Discussion
Dear Peter Vennemann, Please forgive me is the question is not relevant but I quite don't understand the reasoning behind this calculation. Wouldn't a totally white pattern give higher value of correlation than a perfect match?
-
Dear Peter Vennemann, Please forgive me is the question is not pertinent but I quite don't understand the reasoning behind this calculation. Wouldn't a totally white pattern give higher value of correlation than a perfect match?
-
Please forgive me is the question is not pertinent but I quite don't understand the reasoning behind this calculation. Wouldn't a totally white pattern give higher value of correlation than a perfect match?
-
Dear Peter, I still cannot acces your thesis. A message saying "The requested uuid could not be resolved." appears when I click the URL. Does it have any acces restriction? Regards, Sheppard
-
Maria Luisa Moriconi modified a comment on discussion JPIV Open Discussion
Dear Peter, I am working to a paper on the results obtained by the DPIV technique (the base of JPIV if I well understood the 1991 Willert and Gharib paper) applied to the JIRAM images. I have again some doubt on the interrogation windows as used in JPIV. The interrogation windows cover the full image with or without overlapping. If so, how much they overlap? Given that I plan the search domain size basing on the displacements empirically inferred by two successive images, how much other passes with...
-
Dear Peter, as anticipated in my last message, I am glad to share with the community the results obtained by processing JIRAM data with JPIV. These results have been presented during the last EPSC at Riga. Any comments is welcome! Thanks
-
Benjamin Schmidt posted a comment on discussion JPIV Open Discussion
Okay found it. You can open the jvc.-File with a spreadsheet program like excel. The third and fourth column are the x- and y-displacement components.
-
Dear Peter, thanks for clarifying! Regarding the use of cross correlation with cells, I use this to analyze fast but repeated movements. Meaning that my cell population in my analysis doesn't change (same is true for the selected area). So each cell stays more or less in the same place. Thanks again! Regards, Ben
-
Dear Peter, I'm currently studying a rather stationary vector field in a pipe flow and i want to determine the grade of turubulence. For that i need to get the information about every vector in the vetor field. Since saving .jvc-Files as vector files isn't supported at the time, can i print out the information anyhow? I know about printing the data of one profile line into the console, but I need every vector with its horizontal and vertical component. Regards, Ben
-
Tried opening the .jvc file and the vector file is faulty. This is the output upon opening it: The specified file could not be interpreted as vector data. Find attached the display that turns up upon opening the .jvc file
-
Tried opening the .jvc file and the vector file is faulty. This is the output upon opening it: The specified file could not be interpreted as vector data.
-
Peter, If I get time, I will look into implementing this into the JPIV. Hope the math is good. Thanks R Karthikeyan
-
Dear Karthikeyan, well, yes. That would be the maximum possible peak heigt or perfect correlation. Good idea! Regards, Peter
-
Dear Ben, Principally, you can use JPIV (or cross correlation in general) for detecting the movement of any random pattern. Two preconditions must be met: 1. Static background is removed (for example, as described here: https://www.jpiv.vennemann-online.de/howto.html#preProcessing) 2. The patterns are random patterns! This might be a problem using cells. Cross correlation finds similar patterns: This should be the same cell in the second image. But it could be a similar looking cell in the second...
-
Dear Karthikeyan, Try to plot/display the vector file (by double clicking it) to verify, if it is valid. Possibly there is a blank line or so that lets the script crash. Regards, Peter
-
Dear Peter, thanks for sharing your work! It is very nice! I repost this because I accidently posted my comment in the wrong forum. Working as a biologist I'm intrigued by the possibility to use your software for the analysis of cell movement. While conducting some tests I have seen that your software works fine not only for particle based images, but standard light microscopy images (no particles). Are there any problems I could run in ( think publication wise)? or would you recommend to use particles?...
-
Dear Peter, thanks for sharing your work! It is very nice! I repost this because I accidently posted my comment in the wrong forum. Working as a biologist I'm intrigued by the possibility to use your software for the analysis of cell movement. While conducting some tests I have seen that your software works fine not only for particle based images, but standard light microscopy images (no particles). Are there any problems I could run in ( think publication wise)? or would you recommend to use particles?...
-
Dear Peter, thanks for sharing your work! It is very nice! I repost this because I accidently posted my comment in the wrong forum. Working as a biologist I'm intrigued by the possibility to use your software for the analysis of cell movement. While conducting some tests I have seen that your software works fine not only for particle based images, but standard light microscopy images (no particles). Are there any problems I could run in ( think publication wise)? or would you recommend to use particles?...
-
Peter, Thanks for the response. Can we say that for normalisiing the output correlation function, one has to divide by the sum of the squares of the pixel values? i.e. normalised_correlation_map = correlation_function_output/sum(each_pixel^2) ? (forgive my wierd equation) :P Thanks Karthikeyan
-
-
Hi Peter, I tried applying mask but this is the error I am encountering: matlab = "D:\Program Files\MATLAB\R2016a\bin\matlab" Specify mask image, please. Specify output filename, please. "D:\Program Files\MATLAB\R2016a\bin\matlab" -nosplash -nodisplay -r "addpath('E:\Karthikeyan\EXPT\PIV\_PIV processing softwre\jpivlib\matlab\');mask( 'E:\Karthikeyan\EXPT\Burner\PIV\TestSession\testsession01.jvc', 'E:\Karthikeyan\EXPT\Burner\PIV\TestSession\Mask1.jvc', 'E:\Karthikeyan\EXPT\Burner\PIV\TestSession\Mask1.tif');exit;"...
-
Dear Peter, I have been trying to apply the mask but this is the error I am getting upon applying mask and trying to open the result file. matlab = "D:\Program Files\MATLAB\R2016a\bin\matlab" Specify mask image, please. Specify output filename, please. "D:\Program Files\MATLAB\R2016a\bin\matlab" -nosplash -nodisplay -r "addpath('E:\Karthikeyan\EXPT\PIV\_PIV processing softwre\jpivlib\matlab\');mask( 'E:\Karthikeyan\EXPT\Burner\PIV\TestSession\testsession01.jvc', 'E:\Karthikeyan\EXPT\Burner\PIV\TestSession\Mask1.jvc',...
-
Dear Jasmin, The first vector is plotted in the middle of the interrogation window of the fist pass. The fringe at the left and top side is thus half the size of the interrogation window size of the first pass. The fringe on the right side and the bottom is usually a bit wider: It is half the width or heigth of the interrogation window plus the rest of dividing the image width (or heigth) by the number of interrogation windows that fit horizontally or vertically, respectively. I hope, this answers...
-
Dear karthikeyan, For understanding the correlation map, think of the the interrogation area of the second image lying on top of the interrogation area of the first image. Now you multiply the grey values of each pixel pair lying on top of each other. Finally you sum up all products. This gives you one single value of the correlation map. You now move one of the interrogation areas one row or column further and again sum up all products. The value is stored in the correlation map next to the preveous...
-
Hi, What could be the meaning of the value of the exported cross correlation function? I am wondering how to scale it in the range of 0-1, as in 0 is no correlation and 1 is perfect correlation? How to normalise across the vector mapping too?
-
Jasmin Samaras posted a comment on discussion JPIV Open Discussion
Hi there, I am a very new user to PIV and JPIV so I do apologise in advance for any obvious questions! I am processing my images and I notice that the area where in the end the vector fields are plotted does not cover the whole area of my image. I am unsure how to upload pictures to this thread to help illustrate my question. Basically after processing using JPIV the vector plots (when I overlay to the orginal image) miss out all the outer edges, what I am interested in is the bottom edge. I have...
-
Dear Sheppard, My thesis should be available at http://resolver.tudelft.nl/uuid:28a0bbb8-e671-4806-ba52-cca2e6db0d50. 1. Specify the distance between the scan planes (in pixel), like explained here: http://jpiv.vennemann-online.de/settings_vectorprocessing_3drecon.html 2. Load your scan planes (evaluated and post processed vector fields) into the files frame at the right. 3. From the drop-down menu »Vector« select »reconstruct third velocity component« (last item) Does this answer your question?...
-
I wanted to create 3D flow images, and read the question about that and its answer. Now, I cant find the fifth cloumn of my vector images. Woud you show me the process for 3D flow in detail? Besides, why your thesis cant be read now? Sincerely yours, Sheppard
-
Nice to hear, that you have the possibility to present a poster and to cite JPIV! I am very curious to see the results! Success with your presentation!
-
Dear Peter, your software is cited in a poster contribution that will be presented to the EPSC in September. I am reading Dr Westerweel PhD thesis searching for some clarifying about the algorithm used in JPIV software because I must confess to have at today again some doubts on the windows definition and using, tied to some vortices position errors in my results. After the congress I will share the poster with this community. Cheers
-
Dear Alon, You are completely right: I uploaded an old version. I fixed that. The...
-
Alon Oyler-Yaniv posted a comment on discussion JPIV Open Discussion
Hi all, I've been trying to integrate JPIV into a batch processing pipeline for cell...
-
Dear Peter, thank you very much for your very clear explanations. In the past days...
-
Dear Eric, JPIV tries to save the file as »"1b.tiff" "2b1.jvc«. What you want is:...
-
Dear Harris, The JPIV output files (.jvc) are just normal space delimited utf8 text...
-
Once did something similar just for fun (see attachement). You will do better!
-
Dear Maria Luisa, interrogation window size: Within this fragment of the image, JPIV...
-
I am using JPIV to study the flow field of the Jupiter cloud pattern at 5 micron...
-
Harris Anderson posted a comment on discussion JPIV Open Discussion
Does anyone know a simple way that I can separate the horizontal and vertical vectors...
-
Dear Anthony, I notice this kind of behavior, when running in a 32 bit JRE on a 64...
-
Olivier SIRE posted a comment on discussion JPIV Open Discussion
Hello, I am working on the relation between bacterial swarming and local visco-elastic...
-
Megan posted a comment on discussion JPIV Open Discussion
The fifth column of data in the exported file is the peak height of the interrogations,...
-
Javier posted a comment on discussion JPIV Open Discussion
Hi there! I am working on an experiment where 3 or 4 cameras can take pictures of...
-
eric posted a comment on discussion JPIV Open Discussion
Dear Peter, I am trying to test your software on a pair of 2 images (in separate...
-
Anthony LB modified a comment on discussion JPIV Open Discussion
Hi guys, I've just started using this software a few days ago & appreciate it, thanks....
-
Hi guys, I've just started using this software a few days ago & appreciate it, thanks....
-
Hi guys, I've just started using this software a few days ago & appreciate it, thanks....
-
Enrico Turco posted a comment on discussion JPIV Open Discussion
Dear Peter, I have a problem with this script. It stops after exports the first vector...
-
Colin posted a comment on discussion JPIV Open Discussion
Hi all, It seems that i have experienced similar problem too. I have operated the...
-
Dear Arien, there are multiple possible sources for such an error. 1. It might be...
-
Dear Enrico, I assume, you refer to the profile function. Please verify, that »Print...
-
Arien posted a comment on discussion JPIV Open Discussion
Experiencing what appears to be the same problem in Windows 7 Professional N. Is...
-
Dear Peter, I have estimated the flow rate, drawing it in the vector plot. I would...
-
Quick update: using cpulimit to limit Java to 50% CPU. Should provide a better indicator...
-
Quick update: using cpulimit to limit Java to 50% CPU. Should provide a better indicator...
-
Dear all, I have just installed Ubuntu 12.04 LTS 64 bit on my MacBook Pro, and I...
-
Dear all, I have just installed Ubuntu 12.04 LTS 64 bit on my MacBook Pro, and I...
1 >
