Remote Sensing Identification Of Stellera Chamaejasme L.based On Plant Community Ranking And Classification

Grassland is an important part of terrestrial ecosystem structure.Qinghai Tibet Plateau is rich in grassland resources.Its unique alpine meadow ecosystem is very representative in the world.However,under the joint influence of climate and human activities,alpine meadow degradation is serious,and various poisonous weeds spread,which has a great impact on regional economic development and ecological security.Stellera chamaejasme L.has become one of the most serious poisonous weeds in Qinghai Tibet Plateau,and is also an indicator species of Alpine Meadow Degradation.Most of the plant remote sensing recognition methods only rely on spectral differences to judge and classify,ignoring the continuous changes of plant communities with the change of environment in the natural state,and the classification results are easy to cause errors.Therefore,remote sensing monitoring and identification of Stellera chamaejasme L.based on plant community to obtain the classification results with clear ecological significance is of great significance to curb its expansion and spread and maintain the alpine meadow ecosystem.In this paper,the Typical Alpine degraded meadow in the middle section of Qilian Mountains is selected as the research area,and the composition and distribution of Alpine degraded meadow plant community are analyzed by using spectral analysis technology and sorting and classification methods.The spectral difference of Alpine degraded meadow community is discussed by using JM spectral distance.Combined with the coverage of Stellera chamaejasme L.community,the remote sensing identification scheme based on plant community is proposed,Random forest and maximum likelihood classifier were used to further compare the applicability of different schemes for monitoring and identifying Stellera chamaejasme L.community in Alpine degraded meadow.The main conclusions are as follows:(1)The spectrum collected by leaf clip showed that there was significant difference between Stellera chamaejasme L.and other green plants.The redundancy analysis of 11 characteristic bands of Stellera chamaejasme L.and plant data in the study area could better reflect the composition and change of plant community in the study area.The results show that the identification band of Stellera chamaejasme L.white flower has a certain application potential in the study of distinguishing Stellera chamaejasme L.community in Alpine degraded meadow(2)TWINSPAN classification results showed that Stellera chamaejasme L.had become the dominant species of plant community.JM distance calculation of measured spectra of different plant communities in Alpine degraded meadow showed that the spectral separability of plant communities in the study area was low.Combined with the adjusted TWINSPAN classification scheme of Stellera chamaejasme L.community coverage,the spatial distribution characteristics of Stellera chamaejasme L.community with different coverage can be objectively and accurately described(3)The maximum likelihood method and random forest method are used to classify the resonon images in the study area,and the high spatial resolution digital orthophoto images in the study area are used as the true value to evaluate the accuracy of the classification results.The results show that the random forest method has achieved good classification results in the remote sensing identification of Stellera chamaejasme L.based on community,with the highest classification accuracy of 91.06% and kappa coefficient of 0.8556.Compared with the maximum likelihood classification method,it is more suitable for remote sensing monitoring and identification of Stellera chamaejasme L.community in Alpine degraded meadow.TWINSPAN classification scheme combined with Stellera chamaejasme L.community coverage can improve the classification accuracy of three Stellera chamaejasme L.communities in the experimental area,which can be used as an accurate and effective remote sensing recognition and classification scheme,and provide reference for the construction of Stellera chamaejasme L.remote sensing recognition and classification method system...