Study On Assessing The Effect Of Architectural Shadow On Urban Vegetation From Remote Sensing Imagery

Urban green population refers to the artificial planting trees, shrubs, flowers and grass in gardens, street greenbelts, private green space and nurseries, etc. Many studies have Indicated that the green population has the functions of Increasing carbon sink, adjusting regional microclimate, purifying air, beautifying urban environment, etc. therefore being an Important Infrastructure for the maintenance of sustainable development of cities. Collecting the quantitative Information of distribution, abundance and, especially, the annual growth, measured by the net primary productivity (NPP), of urban vegetation is crucially important to the analysis of carbon cycle, to the assessment of vegetation’s benefits to the downtown environment and to the ecological modeling etc. However, changes in environmental conditions will have certain effects on the annual growing rate resulting in obvious disparity of the growing rate even among the plants of the same species and ages. The disparity brings challenges to the fast and automatic estimation of NPP In a large area. Among the factors having influences over the growth rate, the pressure of building shadow is one of the determined and significant factors. Quantifying the shadow pressure and establishing a continuous distribution field of the pressure is crucial for estimating NPP. However, the high heterogeneity of the urban underlying surface increases the difficulty to conduct the quantification of shadow pressure.This paper focuses on three methods, namely, the method of quantifying shadow pressures from remote sensing imagery and establishing their distribution field, the method of regressively analyzing the relationship between shadow pressure and plant growth rate, and the method of predicting slowing rate of plant growth by using the data from the shadow pressure field.(1) The basic method of automatically quantifying shadow pressures and establishing their distribution field can be depicted as:mapping multi-temporal shadow distributions from remote sensing images, getting shadow frequency field from the distributions by image change analysis. Five test images were randomly selected from five groups of false-color aerial near-infrared (nir) images in a time sequence from 1993 to 2006. Five shadow binary images were obtained by supervised classification. The shadow frequency was accumulated from these binary images. in order to separate water body and shadow surface which often had very close features, the feature space for the supervised classification was constructed with several elaborately designed descriptors, including the so-called "density of cut set". The "knowledge and experience and method" was then used for further improvement on the classified water body and shadow surface to obtain accurate shadow binary Images. The accuracy assessment indicates that the accuracy of mapping shadow usually is better than 82%.(2) The basic method of regressively analyzing the relationship between shadow pressures and plant growth rate can be depicted as:taking camphora and cedar as the indicators of broadleaved and coniferous trees respectively, selecting plant samples from remote sensing imagery, representing their annual growth by their changes in crown diameters, and getting the relationship between shadow pressure and plant growth rate by regressive analyses. The R2 of the regression equations of the two kinds of trees are up to 0.97 and the relationships can be trusted.(3) There are several steps to predict the slowing rate of plant growth by using the data from the shadow pressure field:constructing the shadow pressure field of a predicting area according to the method described In section (1), mapping the distributions of conifer and broadleaf plant populations by supervised classification, forecasting the slowing rates of growth of coniferous and broad-leaved plants respectively according to the correlation equations described In section (2) and establishing a field related to the slowing rates. However, it would be necessary to derive the relationship between the slowing rate of plant growth and the relative NPP changes if predicting the impact of shadow pressures on NPP was required. Therefore, the research had to stop in the position before deriving the relationship because there would be a large amount of manpower and material needs to measure NPP In situ for the derivation.Experimental results indicate that the above-mentioned methods are feasible and effective for getting shadow pressure data and measuring the annual growth of sample plants from remote sensing imagery, analyzing the statistical dependence between shadow pressure and plant growth and establishing two continuous distribution fields related to the shadow pressure and the slowing rate of plant growth. The accuracy of supervised classification for separating among different urban common underlying surfaces will be greatly improved by adding the descriptors involved with the density of cut set to the feature space. The method associated with experience knowledge helps to distinguish water body and shadow. Shadow pressure can significantly affect growth of camphor and cedar resulting in slow growth or even stagnation. In addition, camphor often has better shade tolerance than that of cedar.