Fast estimation of stomatal density and stomatal area of plant leaves based on eCognition

ObjectiveLeaf stomatal is a main channel used as exchange matter between plants and environment, which is very sensitive to environmental changes. How to calculate stomatal area and openness data quickly and accurately still lacks mature methods and techniques. This paper aims to explore the quantitative calculation of leaf stomatal density and stomatal area, and provide reference for future research on plant stomatal by this way.

MethodThis study chose the leaf of Fraxinus pennsylvanica, Ailanthus altissima and Sophora japonica as objects, analyzing stomatal information by multi-scale segmentation and classification recognition and classifying the leaf stomatal microscopic images via eCognition image processing software. The stomatal imagines were classified and identified based on the spectral characteristics, bcenterness characteristics and geometric features of the objects.

ResultThe results showed that the best parameters of the stomatal division and the combination of automatic extraction rules were: scale parameters 120-125, shape parameter 0.7, compactness parameter 0.9, bcenterness value 160-220, red light band> 95, shape-density index 1.5-2.2.

ConclusionThe precision of stomatal density and stomatal area extracted by this method was 99.2% and 94.5%, respectively and the results were satisfactory. So the method is suitable for rapid extraction of stomatal information in plant leaves.