Research On On-line Recognition And Concentration Measurement Of Dominant Species Of Algae

The eutrophication degree of freshwater in China is becoming more and more serious,which is resulted in the over-breeding of phytoplankton and the frequent occurrence of algal blooms,which seriously endanger human health and sustainable economic development.However,effective control measures are still lacking.Therefore,early warning of algal blooms is essential.The species and biomass of dominant algae in inland freshwater bodies are closely related to the possibility of blooming and the degree of bloom in the water bodies.The categories and concentrations of dominant algae is needed to be monitored.The categories and concentrations of dominant algae are dynamic process of change affected by season,water temperature,etc.The dominant species of algae in the same water body in different seasons have significant differences,and the dominant species of algae in different waters and depths of the same season also have obvious differences in their categories and densities.Therefore,long-term continuous measurement of water bodies in different positions and depths needs to be carried out.However,there is a lack of methods and devices that can be applied to on-line dominant species identification and concentration measurement.In view of the above problems,based on the study of the phytoplankton fluorescence measurement and the characteristics of phytoplankton photosynthetic pigments,this paper presents a method for on-line identification of dominant algal species based on the fluorescence ratio and a method for the determination of dominant algae concentration based on multiple linear regression.At the same time,an on-line detection device of fluorescence algae based on the above-described methods was developed with four characteristic wavelengths excitation light sources of LEDs.In this paper,the algae species identification method mechanism was deduced based on the characteristics of algae fluorescence ratio.Under the excitation of four characteristic wavelengths,the ratio of the product of fluorescence efficiency,molar absorption coefficient and content of photosynthetic pigment are reflected by the six ratios of pure fluorescence intensity of photosynthetic algae.The photosynthetic pigments in the same categories of algae are similar,but the photosynthetic pigments in different categories of algae have obvious differences.Therefore,the fluorescent ratio of pure algae can be used to reflect the fluorescence ratio of their belonging species.The fluorescence ratio of mixed algae mainly reflects photosynthetic pigment characteristics of dominant algae,but the influence of non-dominant algae fluorescence intensity should not be neglected.Therefore,the recognition of dominant algae in mixed algae based on K-nearest neighbor algorithm and support vector machine class online identification methods are proposed.Based on the identification of dominant species of algae,a method for the classification and measurement of algae concentration is proposed and the effects of scattering and absorption effects in water bodies are corrected by multivariate scatter correction method.In this paper,Chlorella,Microcystis aeruginosa,Phaeodactylum tricornutum,Spirulina,Anabaena and Andocollare are selected for experiments by fluorescence spectroscopy.Classification model is obtained by training samples and the K-nearest neighbor algorithm and support vector machine.The effect of classification and concentration measurement error are discussed through the test samples.The experimental results of 126 samples show that the recognition accuracy of the test based on K-nearest neighbor method is above 97.6%and support vector machine is above 99.2%,and the error of concentration measurement of dominant algae is much lower than that of non-dominant algae.In this paper,an on-line algae detection device based on the fluorescence ratio is developed.Based on four LEDs and detection modules,the device can be placed in a mobile water equipment to realize the dynamic measurement of the water.The device performance and measurement results were tested by experiments.The experimental results show that the performance of the light source module and the detection module of the device are stable,the identification results of the 12 collected samples in the field are accurate and the error of concentration measurement results are 10.86%.