Research On Monitoring Method Of Urban Vegetation Water Content Based On Image

With the continuous advancement of the construction of ecological civilization in my country,the development of urban ecological civilization has received more and more attention.Among them,urban vegetation is part of the construction of green development,and its deployment is an important indicator of urban ecological civilization.The monitoring of vegetation growth is the basis for the sustainable and healthy development of urban greening.Water,as an indispensable condition for the normal growth of vegetation,plays an important role in the normal growth of vegetation.At present,in practice,the irrigation of urban vegetation generally relies on manual timing and fixed-point treatment,and there is a lack of scientific and standardized management of urban vegetation growth.At this stage,the identification of vegetation moisture content is mainly to establish a mapping relationship between specific wavebands and vegetation moisture content.This method requires the help of multi-spectral cameras,which is costly and difficult to promote on a large scale,which does not meet actual needs.Therefore,this paper studies the inversion model of vegetation moisture content based on UAV visible light,develops a prototype urban vegetation monitoring system,and constructs a vegetation growth distribution map,which lays the foundation for the scientific management of urban vegetation,and is useful for future ecological environment assessment and overall urban planning.Far-reaching significance.The specific content is as follows:(1)Aiming at the problem of the lack of visible light urban vegetation images and moisture data sets by drones,six common urban vegetations in the north were selected,and the drone aerial photography ground vegetation observation plan and ground vegetation water content collection standards were formulated to establish a city in a natural environment.Vegetation data set,and use histogram equalization to the collected original vegetation images and enhance the image contrast based on the object Log transformation,which lays the foundation for effectively separating the image vegetation information.(2)Aiming at the problem of clutter and strong interference in the vegetation aerial images collected by drones,this paper proposes a color feature extraction method based on the vegetation index GVSI(Green Vegetation Segmentation Index),and combines the texture features of the vegetation on the image,combined with the threshold The vegetation information is extracted by segmentation,so that the vegetation part and the non-vegetation part on the image are effectively separated.Compared with other color feature methods,the pixel accuracy has been significantly improved.(3)Aiming at the high cost and poor applicability of traditional hyperspectral retrieval of vegetation water content,this paper proposes a visible light vegetation water content retrieval model based on HSCNN(CNN Based Hyperspectral Image Recovery from Spectrally Undersampled Projections).The model uses HSCNN technology to reconstruct the RGB image to restore its lost spectral features,extracts the restored spectral features of the image through the third-order color moments,and then uses regression methods to establish the mapping relationship between the spectral features and the actual vegetation water content to achieve urban vegetation Automatic recognition of water content.Compared with the unreconstructed RGB image and other models,the accuracy of vegetation water content inversion is improved.(4)The prototype system of urban water content monitoring was designed and implemented.The system is divided into five parts: login management module,image management module,vegetation ecological growth distribution map module,and observation point vegetation growth record module.The city is established based on the geographic location information of the image.The distribution map of vegetation growth has played an active role in the scientific management of urban vegetation and the realization of automated urban vegetation monitoring.