Detecting textured objects using convex hull

摘要: In this paper, we present a methodology of locating 3D objects known shapes from single gray-scale image, in particular with rich textures on the surface. While traditional approaches identify by grouping and matching local features, locate object image using its convex hull, high level feature not given much attention literature. A “direct line detection” algorithm is developed to detect segments directly divided small blocks. Lines are clustered hull or group clusters computed edited extract 2D contour object. Successful experiments rectangular boxes cylinders show effectiveness approach potential usage industrial applications.