FIJI/ImageJ Particle Analysis

Hi everyone! I'm Helen, re:3D's newest hire. I've been testing out various printing materials on the GBX, ranging from pellets to plastic flake shredded straight from our grinder.

Recently I've been doing particle analysis on those materials to figure out the optimal particle size for use with the GBX. I still need to do more testing, but particles ~3 mm in length seems to be ideal. Somewhere around 5 mm is when you start running into feeding issues.

For people grinding their own recycled plastic, here is my method for analyzing the particle size of the resulting ground-up particles. Anyone can use this procedure to analyze whatever ground plastic they want, and determine if it's the right size to print with the GBX.

I used the free, open source software FIJI (FIJI Is Just ImageJ) to do the analysis. FIJI can be downloaded here.

I referenced this particle counting tutorial to create the below procedure. 

Procedure:

1. Prepare a black surface with even lighting. Place a 2” white piece of paper on it for calibration purposes.
2. Spread the flake/pellets onto the surface in a single layer.
3. Take a picture and move it to your computer
4. Open Fiji
5. File > Open and choose the picture.
6. Go to Image > Type > 8-bit to convert to black and white
7. Go to Image > Adjust > Threshold and drag the sliders. This converts pixels below the threshold to black and above the threshold to white, resulting in a binary image.
8. With the rectangle selection tool, select the calibration paper. Record the pixel length.
9. With the rectangle selection tool, select one of the smallest particles to measure. Multiply the pixel width and height and round down to the nearest order of magnitude to obtain the minimum pixel area.
10. With the rectangle selection tool, select one of the largest particles to measure. Multiply the pixel width and height and round up to the next order of magnitude to obtain the maximum pixel area.
11. Go to Analyze > Analyze Particles to open the Analyze Particles window.
12. Type in the minimum and maximum pixel sizes
13. Set Show: Outlines
14. Check Display results, Clear results, Exclude on edges, and Include holes. Click OK.
15. A window with outlines should appear. Compare to the thresholded image to confirm good particle detection.
16. A results window should also appear. Go to Results > Distribution… then click OK. A graph of the results should appear. Gut check the results for the correct size range. This sample had a ton of tiny, dust-sized particles.
17. In the results window, go to File > Save As and save as a csv file.
18. Import the csv file into Excel or Google Sheets. Use the calibration value from step 8 to convert the pixels^2 values from the csv file to mm^2 by using this formula:

(area in pixels^2) * [ [ (2*25.4) / calibration value ] ^ 2 ] = Area in mm^2

Now you can compute averages and standard deviations, and make charts on the data to get an idea of the particle size and its distribution. And... ultimately determine if it's viable to print with GBX!

Some caveats: This data only shows the cross-sectional area of the particles and says nothing about their height. These particles are 3D objects, after all. Also, this procedure doesn't collect any data on the shape of the particles (long and skinny vs round, for instance). Next step for me would be to investigate using FIJI to measure sphericity.