Application of image analysis for grass tillering determination

Area of Science:

• Agricultural Science
• Plant Biology
• Computational Biology

Background:

• Tiller number is a critical parameter in grass ecology and breeding.
• Manual tiller counting is a laborious and time-consuming process.
• A need exists for efficient and accurate tiller quantification methods.

Purpose of the Study:

• To develop and evaluate an automated image analysis method for counting grass tillers.
• To assess the applicability of the method across diverse grass species.
• To provide a faster alternative to traditional manual counting techniques.

Main Methods:

• Grass tillers were prepared using cutting and tip painting.
• Images of grass bunches were acquired.
• An automated image analysis procedure segmented shoots based on morphological parameters.

Main Results:

• The image analysis method demonstrated high speed and accuracy in counting tillers.
• Individual morphological parameter sets were required for optimal recognition across different grass species.
• The method was validated on five species: Phleum pratense, Lolium perenne, Dactylis glomerata, Festuca pratensis, and Bromus unioloides.

Conclusions:

• Automated image analysis provides a significant improvement over manual tiller counting.
• This method is suitable for accelerating grass breeding programs.
• Species-specific parameter optimization is essential for robust tiller quantification.