Quantitative image-based assessment of biofilm-induced surface stabilization in saline soils
Year: 2026
Type: article
Abstract: Soil degradation and erosion are major threats to agricultural productivity and ecosystem stability in arid environments, yet the early-stage stabilization of saline soil surfaces by biocrust-forming cyanobacteria has rarely been quantified using a non-destructive approach. To address this gap, we developed a quantitative image-based framework to track micro-topographical changes in field plots inoculated with native cyanobacteria under two salinity levels, low salinity (EC = 6.5 dS m −1 ) and high salinity (EC = 15.4 dS m −1 ), using a completely randomized design with two independent replicates per treatment. After 50 days, standardized digital images of the soil surface were analyzed in MATLAB through grayscale conversion and local texture processing, including pixel standard deviation (SD), image entropy, and moving-window SD/entropy filtering, to capture spatial heterogeneity associated with roughness, cracks, and surface disorder. The inoculated plots showed visibly smoother and more cohesive surfaces than the controls, with significant reductions in pixel variability (14% and 37.36% under low and high salinity, respectively; P < 0.01), whereas entropy decreased by 9.03% and 18.04% without reaching statistical significance (P > 0.01). These results indicate that cyanobacterial inoculation can improve soil surface organization, but its effectiveness declines under stronger salinity stress, likely because osmotic and ionic constraints suppress metabolic activity, limit extracellular polymeric substance secretion, and restrict formation of the filamentous network needed for particle binding and crust development. The study demonstrates that non-destructive image analysis can sensitively detect early biocrust formation and that cyanobacterial inoculation is a promising bioengineering strategy for stabilizing saline soil surfaces when salinity does not exceed physiological tolerance thresholds.
Authors: Shuaili Lv, Shenling Peng, Xiang Zhang
Language: English
Domain: Life Sciences
Open Access status: gold
Estimated APC paid: $3,801
OpenAlex ID: W7225782628