SPATIAL MULTI-SCALE FRACTAL ANALYSIS OF TOMATO LEAF STRUCTURAL DEGRADATION UNDER DISEASE PROGRESSION

Abstract

Tomato (Solanum lycopersicum) is an important horticultural crop worldwide, and its productivity is significantly limited by the widespread occurrence of foliar diseases that gradually modify leaf morphology and physiological function. Disease progression in the leaves entails intricate spatial transformations like lesion development, vein proliferation and disruption in venation, tissue necrosis, and disruption of normal laminar architecture. Often, these structural alterations may be challenging to express in a conventional visual or categorical scoring system, which is often subjective and lacks granularity. In the present study, a spatial multi-scale fractal analysis framework has been proposed to quantitatively capture structural degeneration in tomato leaves during disease progression. Leaf images from sources with known disease severity levels were analyzed through fractal geometry-based metrics, focusing on two key descriptors: fractal dimension (FD) and lacunarity.to express structural complexity and heterogeneity. Fractal dimension serves as a description of overall morphological complexity in this context, whereas lacunarity is widely used to quantify spatial irregularity and gap distributions within leaf tissues. Multi-scale analysis was employed to examine structural changes at different spatial resolutions. This helped us better understand both global shape changes and localized tissue damage. The findings demonstrate a steady reduction in fractal dimension as disease severity escalates, indicating a gradual deterioration of structural complexity. On the other hand, lacunarity values increase as the disease gets worse, which means that the leaf architecture becomes more varied and broken up. The results show that fractal-based spatial analysis is a strong, non-destructive, and objective way to measure how disease damages the structure of plant leaves. This technique holds considerable promise for utilization in automated plant disease diagnostic systems, precision agriculture, and digital plant phenotyping platforms.

Plant phenotyping relevance

病害進行に伴う葉の構造劣化を、画像からフラクタル次元とラacunarityで定量化する解析手法が研究の中心であり、植物表現型の取得・抽出に該当する。

Code and data availability

The paper analyzes tomato leaf images from the public PlantVillage dataset and describes Python/Fiji-based fractal analysis, but no author-deposited dataset, code, models, or supplement with a public URL is provided. The only URLs in the article are the journal website and the CC BY license link, neither of which hosts

No evidence-backed public reproduction asset is currently recorded.