Based on Image Processing Technology Estimatingleaves Stomatal Density of Populus Euphratica and Analysis of Its Ecological Significance

JIAN Shengqi,ZHAO Chuanyan,ZHAO Yang,PENG Shouzhang,PENG Huanhua

2011-01-01

Abstract:Riparian ecosystem plays an important role in maintaining water quality,geomorphology,biodiversity,and aesthetics of the landscape in arid and semi-arid areas.Populus euphratica is a dominant species of the riparian vegetation in Ejina desert located northwest China.Therefore,much more attention has been paid to the species about its growth,distribution and succession.Because of growing in arid area,Populus euphratica is extremely dependent on ambient ground water which is replenished by Heihe River,the second largest inland river in northwest China.Within riparian corridors,ground water varies along lateral gradients.That is,the ground water level typically falls with increasing distance from the active channel due to paralleling increases in flood-plain elevation that result from sediment aggradations.Along the same gradient,water stress for plant growth may increase,because of increase in ground water depth and decrease in replenishment of shallow soil moisture from overbank floods.Accordingly,many ecological and physiological adaptive characteristics of vegetation have been exhibited.The responses of plant leaf physiological status to water stress have been reported,especially the response of leaf stomatal density under water stress.However,little is known concerning the relationships between stomatal density of Populus euphratica and water stress status.Therefore,our objectives are(1) to develop an effective method to obtain stomatal density;(2) to analyze the relationship between leaf stomatal density of Populus euphratica and the status of water stress.Usually,traditional method used to extract the information of stoma from leaves is inefficient.We developed one new method to get stomatal density in the study.The steps are as follows: First,images of stoma were obtained by microscope(Leica DM6000 B).Then,classification of images was performed by the eCognition software.We used a classification technology of the object-oriented.The method has an advantage of auto-identify classification for high resolution image whose information is abundant in structure and geometry.Parameters such as scale parameter,shape and compactness were used to classify in the study.They were determined in trail and error.Scale parameter,shape and compactness were selected to be 230,0.7 and 0.9 respectively.And then,the classified images based on object-oriented classification method were imported into ArcGIS for calculating stomata pixel value and cells of standard stomata,which are used to calculate the number of stomata on one image.Finally,stomatal density can be obtained by dividing the stomata number with the image area.The 18 sampling plots were selected in the study area(the lower reaches of Heihe River).5400 leaves stomatal images were obtained.A batch program was made using R language code to deal with these images based on the steps mentioned above.The results show that: the method in the study has high efficiency and accuracy to obtain leaves stomatal density.There are variation of stomatal density in 18 sampling plots,ranging from 76.7 stoma/mm2 to 139.4 stoma/mm2,average value 105 stoma/mm2.With the increase of water stress,the stomatal density expresses fluctuant change,from decrease to increase to decrease.