Research Of Assembly Line Model Matching Based On Random Forest And Geometric Hash

Assembly line production has been playing a dramatically importantrole in industrial production, specifically in the present automationproduction process. It will reduce the production and maintenance costsgreatly and also improve the efficiency of automatic production if theobjects in the assembly line can be classified automatically. However, it isa very important but also a tough issue to match objects in the field ofpattern recognition as objects may be influenced by themselves such asscale, rotation, obstacle and also surrounded environment such asintensity change in the process of recognition. Therefore it is a toughissue to match objects in a complex environment. In order to solve thisproblem, Random Forest algorithm as well as Geometric Hashing methodis introduced cooperated with feature extraction of key points. Where,Random Forest can construct clarifier to match models by machinelearning. Different categories of training features are extracted to optimize the matching result; While Geometric Hashing methodcombines structural information with hash table to match models. Theimprovement of Geometric Hashing algorithm is proposed to match moremodels rather than only one model and also the accuracy and speed areincreased.