Rapid algal culture diagnostics for open ponds using multispectral image analysis

This article presents a multispectral image analysis approach for probing the spectral backscattered irradiance from algal cultures. It was demonstrated how this spectral information can be used to measure algal biomass concentration, detect invasive species, and monitor culture health in real time. To accomplish this, a conventional RGB camera was used as a three band photodetector for imaging cultures of the green alga Chlorella sp. and the cyanobacterium Anabaena variabilis. A novel floating reference platform was placed in the culture, which enhanced the sensitivity of image color intensity to biomass concentration. Correlations were generated between the RGB color vector of culture images and the biomass concentrations for monocultures of each strain. These correlations predicted the biomass concentrations of independently prepared cultures with average errors of 22 and 14%, respectively. Moreover, the difference in spectral signatures between the two strains was exploited to detect the invasion of Chlorella sp. cultures by A. variabilis. Invasion was successfully detected for A. variabilis to Chlorella sp. mass ratios as small as 0.08. Finally, a method was presented for using multispectral imaging to detect thermal stress in A. variabilis. These methods can be extended to field applications to provide delay free process control feedback for efficient operation of large scale algae cultivation systems.