Using computer vision to monitor germination time course of sunflower (Helianthus annuus L.) seeds

Abstract

Summary A computer-controlled system was used to monitor the germination time course of sunflower seeds. The system integrates a Jacobsen table, a controlled lighting environment, a video camera and a computer. Software was developed to control image recording and image analysis. The algorithm retained for detecting and counting the germinated seeds was validated in two ways: 1. First, human operators read images independently and the variability between readers was compared to the computer reading. 2. Then, computer counts were compared to counts of germination carried out directly on the Jacobsen table. The whole equipment (automatic shooting and algorithm connected to the germination system) was tested to plot germination time courses of three sunflower seed lots at 20°C. Detailed germination curves were obtained allowing a perfect fitting in a probit model and a comparison of the seed lots. These results confirm the high potential of artificial vision in quality evaluation of seeds.