A Study On Quantity And Sampling Optimization Of Wetland Plant Species Diversity Based On UAVRS

With the increasing use of wetland resources influenced by human economic activities,some wetland resources have been damaged severely,and the diversity of wetland plant species has become a hot issue around the world.As is known to us all,species diversity information is the basis of the wetland resources research and achieving the ecological and socio-economic function of wetland.However,due to the complexity of the wetland ecosystem,the limitation of data mining tools and the subjectivity of basic information collection methods,wetland resources research still lacks high-quality basic data.Therefore,in order to ensure that the sampling results can not only reflect the overall characteristics of the wetland plant community,but also save manpower,material financial and time resources,it is essential to choose an appropriate sampling method.In this paper,an“association" which is a quadrilateral with lengths of 158m,145m,152m,and 162m on each side,was selected and aerially photographed with Phantom 4 Pro in August 2019 for the research in the Yangzizhou Wetland in Nanchang.The aircraft has flown four times in total and the flight height was 10 meters.With spatial resolution being 3mm,1477 photographs were obtained.The Pix4D software was used in a graphic workstation to generate an orthophoto map of the wetland plant communities in the research area,then the area of the space sampling box was redefined with 4304m2 in practical calculation.Based on the plant information stem from human's investigation and collection,combined with the shape,texture,position and color features of the visible light image,the obtained actual UVARS image was visually interpreted to obtained different plant species vector map layers.15 plants which belong to 12 families and 14 genera,and are dominated by perennial herbs were investigated in the research area.Among them,alternanthera philoxeroides and cynodon dactylon are the dominant species,xanthium strumarium and polygonum hydropiper are common species.The frequency of most species is less than 0.1 and the important value is low.Based on vector map layers of 15 plant species,the Shannon-Wiener Species Diversity Index was calculated by using a grid of lm*1m(the usual sample size used on the wetland sampling research)and 4304 rasters or samples were obtained totally.The Shannon-Wiener index of each sample was calculated to be between 0 and 1.301,indicating that the plant species richness is relatively low and single in the experimental area.The parameter estimation results of Shannon-Wiener index obtained from sampling,Systematic sampling and subjective sampling are made under the precondition that the exact value of species diversity index(the mean value of the Shannon-winner index of all sampling units in the space sampling box)and the confidence level is made at 95%,while the allowable error is made at 15%.The absolute errors of simple random sampling,systematic sampling,and subjective sampling were 0.0217,0.0253 and 0.0135,respectively;the relative errors were 15.32%,14.17%and 14.35%respectively.The results show that the absolute error difference of random sampling is the largest without the mean value treatment,which indicates that the method is less stable in wetland plant species diversity sampling,and the relative error is slightly larger and therefore it is not recommended to apply directly to species diversity research.The overall sampling effect of systematic sampling is relatively desirable with its stability is relatively high,the relative errors is the smallest and the absolute errors fluctuates steadily.The results of subjective sampling are better than random sampling and systematic sampling.After averaging the sample parameters obtained by the three sampling methods,it is obvious that the subjective has an advantage in species diversity study with its absolute error is only 0.0135,which indicates the obtained Shannon-Wiener index is closest to the theoretical value.This paper shows that using UAVRS can help obtain high-quality,high-resolution species diversity data,and it is both time-saving and labor-saving;Besides,using subjective sampling can significantly increase species diversity sampling survey accuracy comprared with random sampling and systematic sampling,especially when investigators are experienced.