Digital image processing method for estimating leaf length and width tested using kiwifruit leaves (Actinidia...

Area of Science:

• Agricultural Engineering
• Plant Science
• Computer Vision

Background:

• Accurate plant leaf dimension measurement is crucial for agricultural research and yield prediction.
• Traditional methods can be labor-intensive and less precise, especially for leaves with irregular shapes.

Purpose of the Study:

• To develop and validate an image-analysis-based method for estimating kiwifruit leaf length and width.
• To assess the accuracy and performance of the proposed method compared to existing approaches.

Main Methods:

• Utilized image segmentation and correction to obtain orthophotographs of kiwifruit leaves.
• Extracted pixel coordinates from leaf edges to calculate length (tip to petiole) and width (widest region).
• Evaluated results using Mean Absolute Error (MAE), Root Mean Square Error (RMSE), and accuracy.

Main Results:

• The method achieved high accuracy rates of 98.88% for length and 98.83% for width.
• Maximum MAE and RMSE were 0.281 cm and 0.366 cm, respectively, across three cultivars.
• Demonstrated superior performance compared to other prevalent measurement approaches.

Conclusions:

• The proposed image-analysis method accurately determines kiwifruit leaf dimensions, including those with concave tips and petiole insertions.
• The method's low error rates and high accuracy validate its effectiveness.
• Potential for extension to other plant species with comparable leaf morphology.