Research On Growth Characteristics Of Microalgae Biofilm Based On Hyperspectral Spectroscopy

The efficient cultivation of microalgae is the prerequisite for the development and utilization of bio-energy.At present,there are two ways to culture microalgae:suspension culture and immobilization culture.Although the microalgae suspended culture mode is easy to operate,the recovery is difficult and complex,and the cost of harvesting is relatively high.The microalgae immobilized culture has many advantages: mode is stable,the biomass is high,and harvest is convenient and quick,and so on.It has become the focus of the researchers in the field of microalgae.But the quality of microalgae,especially microalgae oil content,is affected by biofilm growth factors significantly.Therefore,it is very important to develop a fast and simple on-line detection method to monitor the growth information of microalgae biofilm in order to adjust the growth condition in time to improve microalgae biofilm.Hyperspectral imaging technology is a new non-destructive,fast and accurate detection technology.Combining the advantages of computer vision and spectral detection,the technology can record the abundant information of biofilm growth well.It has found a simple and effective method for rapid nondestructive detection of microalgae biomass.In this paper,the hyperspectral technique and biometric technique were used to analyze the characteristics of microflora reflectance spectra and hyperspectral images under different growth conditions.A new method for detecting microbial biomass of microalgae was proposed,which provided a technical support for the rapid acquisition of biofilm growth information and the efficient cultivation of microalgae biofilm.The results of the study are as follows:(1)Culture microfilm biofilm under different temperature,different pH and different light intensity.(2)According to the theoretical basis of hyperspectral imager,a hyperspectral image acquisition system was established,and the growth information of microalgae biofilm was detected.The hyperspectral information of microalgae biofilm under different temperature,pH value and light intensity was obtained.(3)There are three distinct peaks in the original spectral reflectance of the microalgae biofilm,and the spectral reflectance corresponding to the characteristic wavelength(789nm,811 nm,930nm)decreases with the increase of culture time.(4)According to the spectral reflectance data collected from microalgae biofilm,25 spectral characteristic variables related to the microbial biomass of microalgae were extracted by measuring the spectral characteristic variables of biomass,and the correlation coefficient of each characteristic variable was calculated.(5)The spectral characteristic variables of the first four spectral correlations with the dry matter of microalgae were extracted by the correlation value of the correlation coefficient,namely: SDy)+SDy)/(SDr-(SDr,SDr,OSAVI and NDVI.Three kinds of mathematical predictive models were constructed,and the prediction models of BP neural network prediction model,single spectral characteristic variables and multi-feature variable fusion were analyzed and compared.The time and precision of the model were used to measure the results.The results showed that the multi-feature variable fusion model had the highest comprehensive evaluation of microalgae biofilm.(6)Based on the results of the prediction of the original spectral reflectance and the predicted model of the microalgae biofilm,the optimum growth environment for the culture of Chlorella vulgaris is: the temperature is 28 ?,the pH value is 8,the light intensity is 3500 lx,The model can also predict the amount of microalgae biofilm at different times.