The automation of cost estimation for manufacturing processes is a challenging task in computer-aided manufacturing. In this paper, we introduce a two-step approach to estimate the cost of injection molds. In the analogy step, data of molds are partitioned into homogeneous groups, and in the mathematical step regression models within each group are constructed. The regression models within each group are based upon geometric and topological complexity that can be extracted automatically from one orthographic 2D image of the injection-molded part print. Based upon the 2D image, wavelet and regional descriptors of boundaries as well as other inherent shape properties, including size, and number of boundaries, is used to construct regression models. Mean estimates and prediction intervals are calculated on a subset of relevant molds so that we can evaluate the risk of an inaccurate bid.
展开▼