Remote Sensing Retrieval Of Chlorophyll A Concentration In Baiyangdian Lake Based On Semi-supervised Soft Classificatio

Chlorophyll a(Chl-a)is an important indicator to evaluate the primary net productivity and eutrophication degree of water bodies,which is of great significance to the assessment of water environment quality.Different from pelagic water bodies,the optical characteristics of inland water bodies are complex due to the input of a large number of terrestrial materials,which brings some obstacles in the retrieval of chl-a concentration.Therefore,the high-precision inversion of Chl-a concentration in inland water bodies is a hot topic in related research fields.Baiyang Lake is an important part of the ecological civilization construction in Xiong’an New Area,and its water environment monitoring and control is of great significance to the ecological civilization construction of Baiyang Lake.In this paper,a semi-supervised soft classification hybrid model inversion algorithm is proposed for the characteristics of Baiyangdian water body,and the accurate inversion of Chl-a concentration of Baiyangdian water body is realized by using the measured spectral data of water body and the remote sensing data of Zhuhai-1(OHS)hyperspectral satellite respectively,which provides technical support for the solid ecological protection of Baiyangdian.The research content and conclusions of this paper mainly include the following aspects:(1)The semi supervised soft classification method is proposed in this paper based on the study of optical classification of water bodies.Combined with 102 measured spectral data sets of Baiyangdian Lake water bodies,the optical characteristics of Baiyangdian Lake water bodies are analyzed.With reference to the slope algorithm of reflection spectrum,a small amount of water body data are classified and labeled.Using the semi supervised fuzzy clustering method,Baiyangdian Lake water bodies are divided into algae dominated Algal non algal particulate matter co dominated and non algal particulate matter dominated water bodies bringing the chance for subsequent Chl-a concentration inversion.(2)A hybrid model inversion algorithm for Chl-a is proposed in this paper inspired by the soft classification results of the optical properties of water bodies.Through the measured spectral data,the red near red band model and the three band bio optical model are optimized,and the optimal weight between models and the membership degree between water bodies are determined to estimate the Chl-a concentration by weighting.Compared with the non-classification inversion method,the hard classification inversion method and the soft classification single model inversion method,the method proposed in this paper shows good performance.The determination coefficient(R~2)of model accuracy verification reaches 0.80,the average relative error(MRE)is 10.40%,and the model fitting degree is the highest and the error is the smallest.(3)The relevant technical methods were transferred to OHS data to retrieve the concentration of Chl-a in Baiyangdian Lake employing the measured spectral analysis results,and the accuracy was compared with common inversion methods.The results show that the inversion algorithm R~2 of the semi supervised soft classification hybrid model proposed in this paper reaches 0.71,and the MRE is 12.63%,which is obviously superior to other methods,proving that this method can be applied to OHS data.A meaningful attempt has been carried out to explore the Chl-a inversion of the inland water body in Baiyang Lake,bringing a new channel for the Chl-a remote sensing inversion of Baiyang Lake.Furthermore,it has been verified on the measured spectral data and OHS data respectively.