Public Lab Research note

Computer vision microscope observations of Elodea plant cells

by MaggPi | August 14, 2018 03:02 14 Aug 03:02 | #16922 | #16922

MaggPi was awarded the Basic Barnstar by stevie for their work in this research note.

Introduction: This note describes a general computer vision approach for microscopic real time image analysis. Methods used in prior Public Lab research notes such as manual gain control and computer vision color detection have been combined into a single software program and applied to microscopic applications. Project videos can be seen at: and http:/


Project description: Major elements of the system design include: -Raspberry Pi NoIR camera v2 with picamera interface.
-Raspberry PI 3B+: The latest version with improved processing speeds.
-Microscope objective with camera -

-OpenCV - Computer Vision video processing software to display camera settings and images. A quad video format is used to show different image analysis techniques.

-Elodea plant cells – Elodea was selected as an imaging object since it only has 2 cell layers and maintains cytoplasmic movement after removing the leaf from the stem.

Software Display: Software generates a video image display split into four viewing areas: .

-Upper left quadrant: Camera settings are displayed in the upper left quadrant and can be viewed while the user is adjusting camera settings.

-Upper right quadrant: displays standard color image.

-Lower right quadrant: Selects green image colorspace to identify cells. Filter settings H:80-156, S: 80-126. V:144-255.

-Lower right quadrant: Sobel filter used to extract image edges.


Microscope slide preparation: Elodea leaf was removed from stem and placed on a wet mount slide with cover slip (without stain).

Video description: -Video describes 3 minutes of continuous Elodea plant cell observations. Note that the mpeg4 video format reduces the video time from 3 minutes to 1 minute and creates a time lapse effect. Different processing options show different aspects of cell motion. Video is best viewed at full screen. Additional video with added colormap can be viewed at http:/

Final thoughts: While this example is designed for a moving plant cells, the general technique of camera optimization combined with computer vision processing could be applied to other areas as well. Microplastics or other particles could be filtered by color, shape and size as discussed in the following topic areas:

Software code at:

Related posts:

@warren, @icarito, @amirberAgain, @Bronwen


This is really cool. Just a couple thoughts on presentation -- want to try making a Gif from this video? I think it'd be neat for people to see quickly. Also, would you mind embedding your video in the post? Thanks!

Is this a question? Click here to post it to the Questions page.

Reply to this comment...

Real-time sobel! Cool -- the edge detection on #image-sequencer is pretty slow... i wonder if we could get this running faster (up to video speeds) using this OpenCV transpile: -- in Image Sequencer, that is!

Reply to this comment...

@stevie awards a barnstar to MaggPi for their awesome contribution!

Reply to this comment...


Here is a GIF but not certain the link is embedded. @warren

Is this a question? Click here to post it to the Questions page.

Reply to this comment...

Login to comment.