Free and open-source software for object detection, size, and colour determination for use in plant phenotyping

Area of Science:

• Plant science
• Image analysis
• Phenotyping

Background:

• Object detection, size, and color determination are crucial in plant science for phenotypes like fruit ripeness and plant health.
• Existing methods often rely on proprietary software or require coding expertise.
• There is a need for accessible, adaptable tools for plant image analysis.

Purpose of the Study:

• To develop and present a suite of free and open-source Python scripts for plant image analysis.
• To enable accurate object detection, size calibration, and color extraction without code adaptation.
• To demonstrate a non-destructive, cost-effective method for plant phenotyping.

Main Methods:

• Utilized a ColourChecker chart for background and color correction.
• Developed scripts for object identification and size calibration using a known object.
• Extracted average object colors in RGB, Lab, and YUV color spaces.
• Applied exponential models to predict lycopene in tomatoes and chlorophyll in basil.

Main Results:

• Free and open-source Python scripts were created for automated image analysis.
• The scripts successfully performed background and color correction, object detection, and size calibration.
• Accurate prediction of lycopene content in tomatoes and chlorophyll content in basil was achieved.
• The method demonstrated consistency across different lighting conditions and camera types (DSLR, mobile phones).

Conclusions:

• A fast, cheap, non-destructive, and inexpensive method for plant material size and color determination was established.
• The developed Python scripts and hardware rig (lightbox, camera, color checker) provide an accessible solution.
• The method accurately predicted key biochemical compounds (lycopene, chlorophyll) in plant samples.