Chlorella vulgaris FSP-E cultivation in waste molasses: Photo-to-property estimation by artificial intelligence

článek v časopise ve Web of Science, Jimp

en

This progress of industry revolution, which involves reutilizing waste materials and simplifying complex procedures of analysis through artificial intelligent (AI), are the current interest in automated industries. There are two main objectives, firstly, the use of waste molasses from sugar mills as a cultivation medium for microalgae and nutrients extraction. The biomass in 15% of the molasses medium without carbon dioxide aeration during cultivation obtained the highest dry cell weight at 1206.43 mg/L. Protein content in the biomass of 10% molasses cultivation medium is 20.60%, which is higher compared to commercial mediums. Secondly, the exploitation of the deep colouration properties of molasses-cultivated microalgae, a novel photo-to-property estimation was performed by k-Nearest Neighbour (k-NN) algorithm through RGB model pixel raster in the images to rapidly determine the biomass concentration, nitrogen concentration and pH without use of tedious analytical processes. The k-value at 4 was studied in normalized Root-Mean-Square-Error (RMSE) for biomass concentration at 0.10, nitrate at 0.11, and pH at 0.02 for a sequence of days.

This progress of industry revolution, which involves reutilizing waste materials and simplifying complex procedures of analysis through artificial intelligent (AI), are the current interest in automated industries. There are two main objectives, firstly, the use of waste molasses from sugar mills as a cultivation medium for microalgae and nutrients extraction. The biomass in 15% of the molasses medium without carbon dioxide aeration during cultivation obtained the highest dry cell weight at 1206.43 mg/L. Protein content in the biomass of 10% molasses cultivation medium is 20.60%, which is higher compared to commercial mediums. Secondly, the exploitation of the deep colouration properties of molasses-cultivated microalgae, a novel photo-to-property estimation was performed by k-Nearest Neighbour (k-NN) algorithm through RGB model pixel raster in the images to rapidly determine the biomass concentration, nitrogen concentration and pH without use of tedious analytical processes. The k-value at 4 was studied in normalized Root-Mean-Square-Error (RMSE) for biomass concentration at 0.10, nitrate at 0.11, and pH at 0.02 for a sequence of days.

Microalgae, Microalgae cultivation, Chlorella sp., Molasses, Artificial intelligence, Image analyze algorithm

15.12.2020

Elsevier

Oxford, England

1385-8947

402

126230

1–10

10