Study On Growth Model Of Haematococcus Pluvialis

As a freshwater green algae,Haematococcus pluvialis has high economic value.The growth status of Haematococcus pluvialis is easily affected by the surrounding environmental factors.The algae cells have high requirements for the surrounding growth environment,in which temperature,light intensity and p H value are the main and key factors.At present,the culture process of algal cells is mainly divided into cell swimming stage and cell immobility stage.The two growth stages are also called cell proliferation stage and astaxanthin accumulation stage.Because the growth cycle of Haematococcus pluvialis is long and the growth state presents a nonlinear trend,in order to study the cell growth state,the intelligent algorithm has strong function fitting ability,takes the number of cells and cell radius in the cell proliferation stage of Haematococcus pluvialis as the performance index to evaluate the growth state of algal cells,and uses experimental equipment such as cylindrical photobioreactor,The three growth environment parameters were monitored in real time,and the algal liquid was collected regularly every day for detection.The number of cells,cell radius and cell image were collected by microscope and Image View software.The growth prediction model and cell image classification model of Haematococcus pluvialis were established respectively for cell growth prediction and cell image classification.The main work of this paper is as follows.Firstly,the cell characteristics of Haematococcus pluvialis cells in cell proliferation stage and astaxanthin accumulation stage were described,and the development status of Haematococcus pluvialis culture method,time series data prediction and cell image classification were introduced.Secondly,the collected time series data of cell number and cell radius are EMD decomposed by MATLAB software,the decomposed noise modal function is discarded,the signal in the mixed modal function is denoised by wavelet,and finally the time series data after wavelet denoising is superimposed with the information modal function and residual component,That is,the time series data of the reconstructed cell radius and the number of cells are obtained.Thirdly,the growth state prediction model of Haematococcus pluvialis is established.The results of using LSTM algorithm to predict the original time series data are not ideal.The reconstructed time series data are analyzed by using the optimized LSTM algorithm,namely emdwt-lstm algorithm.The results of the two algorithms are compared and analyzed.Finally,it is found that the prediction effect obtained by using the optimized algorithm is better.Fourthly,in the MATLAB software environment,the hog-svm algorithm is used to classify the cell images in the cell proliferation stage and astaxanthin accumulation stage.Firstly,the twostage training is concentrated on the cell image,the image features are extracted by using the directional gradient histogram,and each image is traversed.Then the cell image classification model is established by using the SVM algorithm.Finally,the accuracy of the model is detected by using the ten fold cross validation method to complete the cell image classification.Finally,the hardware environment such as the culture device and data acquisition method of Haematococcus pluvialis are described.After a series of comparative experiments to explore the optimal environmental parameters in the two growth stages of Haematococcus pluvialis,the optimal temperature,light intensity and p H value for algal cell proliferation and astaxanthin accumulation were obtained respectively,and the three environmental parameters were controlled in a photobioreactor.