Automatic Leaf Epidermis Assessment Using Fourier Descriptors in Texture Images

Area of Science:

• Botany
• Computer Science
• Image Analysis

Background:

• Plant species identification is challenging, often requiring expert knowledge and seasonal plant parts like flowers or fruits.
• Leaves are available year-round, making them a consistent source for plant identification.
• Texture image analysis offers a method to extract quantifiable features from leaf images, correlating texture with species.

Purpose of the Study:

• To present a protocol for extracting texture features from microscopic leaf images.
• To assist in the specialized work of plant species identification.
• To enable automated analysis of leaf epidermal images.

Main Methods:

• Leaf epidermal dissociation for sample preparation.
• Microscopic imaging of leaf samples.
• Texture feature extraction using Fourier analysis.
• Development of an automatic method for image analysis.

Main Results:

• Established a relationship between leaf texture and plant species.
• Developed a protocol applicable to microscopic leaf images.
• Demonstrated the potential of texture analysis for species identification.

Conclusions:

• Texture analysis of microscopic leaf images provides valuable data for plant species identification.
• The described protocol facilitates automated analysis, reducing reliance on expert identification.
• This approach offers a consistent method for plant identification using readily available leaf samples.