Fast Detection Technology Based on Machine Vision for Hill-DropPerformance of Precision Seed-Metering Device

摘要

Small seeds such as vegetable seeds and flowerseeds are sowed by hill-drop planter. Based onmachine vision, a vision inspection system of hill-dropperformance of precision seed-metering device wasestablished for small seeds. The seed regions weremarked by using labeling algorithm, and seed featurewas extracted by detecting the location coordinates ofthe seeds. Hill spacing was calculated by measuringseed spacing first and then separating hole, and pointgrains was counted. The experiment showed that theresults measured by the system were in agreement withmanual detection. The hill spacing error within 3 mm,and point grains error less than 25%.