Identification and counting of natural donor blood erythrocytes of human by digital optical microscopy method using spectrally filtered illumination

The possibility of identification and counting of native donor blood erythrocytes based by static, non-flow digital optical microscopy was studied. The object of study was the whole donor blood diluted by saline and placed into Goryaev's counting chamber. The sample was examined in transmitted light by a Lumam P-8 digital optical microscope equipped by a Basler acA920-40um camera. In order to identify erythrocytes by spectral characteristics two sets of micrographs of 20 pieces in each were obtained. In the first set there was no optical filter in the illumination channel of the microscope and in the second set an interference filter for a 420 nm wavelength with a 10 nm bandwidth was used. The characteristics of the interference filter were chosen to come more close to the Soret band of the RBC hemoglobin to obtain the best contrast for photographic images. The technique for automated analysis of micrographs using the OpenCV software was developed to recognize erythrocytes and to carry out their counting. The results of the computer RBC counting were compared with the manually counted red blood cells. It was found that for computer approach accepted the proportion of the erythrocytes identified and counted in averages was 97–98%.