Three Dimensional Seed Shape and Size Measurement with Orthogonal Cameras

Abstract

Seed shape, size, size uniformity and seed coat wrinkles are important factors influencing trade in many grains. Shape of the grain sometimes makes it hard to measure certain seed dimensions accurately at a practically acceptable rate. Lentil plumpness (determined by 3D shape and size characteristics such as seed diameter, thickness, edge curvature, etc) is an important seed characteristic affecting its splitting ability, a property most sought after by the splitters or processors of red lentils that are usually traded after splitting. Physical measurement of lentil shape and size characteristics is monotonous and time consuming. A dual camera system was used to image individual lentil seeds from two angles to determine 3D size and shape characteristics. Image analysis (IA) algorithms were developed to measure seed diameter, thickness and edge curvature. The IA measurements correlated well with the physical size measurements (R2 > 0.95). Results of this study along with the camera setup issues are discussed in this paper.