Research On The Prompt Steel Bar Separation System Based On Multi-view

Regarding the usage of Computer vision system in industrial production line, the difficulty are always real-time, accuracy and robustness of the system As discussed in this article, "Entirely spilt-steel system by visual feedback control" refers to the computer visual system with multi-cameras, being applicable to the steel& iron enterprises with rough and complex working conditions. The system analyzes the distributed images of the deformed steel bar on forwarder, and then controls multi-spilt-steel components to a coordinated actions, achieving the auto-spilt of the around 10m deformed steel bar. The contents are as followed:System design: Analyze all existing splitting-steel system for counting and spilt-steel tips. A program is set to lift the spilt-steel wedge in turn and make the steel-bar separated completely in the situation where one end of the steel-bar has been separated. Propose a visual system design of multilist for subsection detection, and decrease difficulty of the multilist visual system, which is also a good help for the installation and maintenance, increasing the robustness of the system.Analysis of the image data: Including filtering and cleaning up, section seperation of steel-bar, examination and detection of separation opening. According to the specification of the working image, take smoothness and direction smoothness adjustable sequence-quick filtering as a method to filter noise from vertical edges and other directions. Propose a method to apply binaryzation by combining with specifications of image edges.In the situation where light intensity and target gray value varies greatly, it still achieve quick and correct detection. Propose a edge detecting process basing on existing knowledge and experiences. According to the existing motion of spilt-steel bar, it can determine the centerline of the area under detection, and then matches and fits the equation of the straight line which can best indicate and describe the point set by means of Hough Transform. The method is of good robustness, being applicable to the situation with strong noise and discrete frontier point, and it can meet the needs of the system timeliness.Decision-making of multi-spilt-steel: In the framework of multilist visual system, it selects an appropriate technology pathway to optimize overall the multi-spilt-steel wedge. It increases the efficiency when it use less spilt-steel wedge during the process of separation.It designs and achieves a completely imitation of the spilt-steel framework, with basic program debugging and design validation completed in the lab; then it is applied to the systematical engineering basing on the imitation framework.On-site usage indicates that the system structure and algorithm are both of effectiveness and feasibility, and it lays a steady consolidation for the widespread of this new system.