| Electrical resistance tomography(ERT)is a commonly used non-destructive testing method for wood.It is often used to detect the internal defects of living trees because of the advantages of convenient carrying and fast testing.In different seasons,the influence of environmental temperature,water content of trees and other factors on resistance of living trees is obvious.Therefore,it is of certain significance to study the ERT image of healthy and decayed trees of different tree species in different seasons,which can provide reference for health assessment of trees in different seasons and forest management.In this paper,four typical tree species in Northeast China were selected as the test objects,including Salix matsudana,Fraxinus mandshurica,Larix gmelinii,Pinus sylvestris.The environmental temperature and water content of each tree species in different seasons were measured,and the resistance data of trees were obtained by ERT instrument.The maximum and minimum resistance values of all trees species were obtained from ERT image of the standing trees,and the proportion of heartwood,sapwood area and its decayed area was obtained by extracting the color characteristics of the ERT image.Taking the data obtained above as the seasonal variation index of the trees,statistics of seasonal variation of healthy and decayed wood samples from two aspects of electrical resistance and ERT image characteristics were conducted,and the changes of healthy and decayed living trees were compared.The regression models of temperature,moisture content and each index were established to analyze their correlation.The specific analysis and main conclusions of this study are as follows:(1)The seasonal variation of the resistance value of healthy standing trees was more uniform and regular in a seasonal cycle(2017.09?2018.09),showing the trend of increasing from October to December and decreasing from March to May.Most trees appeared step phenomenon near 0?.The whole change range of electrical resistance value was larger than that of healthy tree,and the step temperature range of rotten tree of some tree species was also changed compared with that of healthy tree(2)The seasonal variation of resistance value of living tree was significantly affected by environmental temperature and water content of trees(P<0.01).R2 of binary linear regression model was higher than 0.8,and the fitting condition was good.(3)The decay proportion of heartwood and sapwood obtained by extracting the color feature of ERT images were significantly correlated with the quality loss rate of heartwood and sapwood(P<0.01),which indicated that the decay proportion of heartwood and sapwood were reliable indexes to characterize the rotten degree.(4)The seasonal variation of heartwood and sapwood area proportion of all tree species detected by ERT were uniform in a seasonal cycle(2017.09?2018.09).There were different characteristics in the seasonal variation of the decay proportion of heartwood and sapwood of decayed standing trees.The variation of the decay proportion of all standing trees can be divided into five types.Multiple regression models were established according to the variation of different types of decayed proportion,among which,the regression models of the second and the third types of variation fitted well,the decayed proportion of living trees was significantly affected by the environmental temperature and the moisture content of trees(P<0.01),the independent variables of the second type were positively correlated with the dependent variables,the independent variables of the third type were negatively correlated with the dependent variables.The regressions of first and the fourth type R2 of the regression model was very low,and there was no significant correlation between the independent variable and the dependent variable(P>0.1). |
PDF Full Text Request