Abstract: In order to solve the problems of time-consuming and low accuracy in manual counting during the process of grain breeding, a method of counting using computer processing of grain particle images was proposed. Firstly, the collected grain particle images was binarized to separate the background and target regions in the image; then morphological operations, object segmentation, and other preprocessing on the image were performed to minimize the phenomenon of grain particle adhesion in the image; in the final counting link, in order to improve the counting speed, based on the important feature that the grain particle images had no holes after preprocessing, the image Euler numbers algorithm was used to replace the traditional connected domain marking algorithm to achieve the counting work.The results showed that the counting result of the image Euler numbers algorithm was completely consistent with that of the traditional connected domain labeling algorithm, but the counting speed of the image Euler numbers algorithm was obviously better than that of the traditional connected domain labeling algorithm.

Key words: Euler numbers algorithm, grain particle images, counting method