Target Localization Method Based On Embedded Vision With FPGAs And Its Application

Intelligent manufacturing technology always leads the development of manufacturing automation.Machine vision technology can be used to improve its automation as well as promote intelligent manufacturing.Compare to traditional machine vision solutions,embedded machine vision has the advantages of small size,low power consumption,and good real-time performance.In this thesis,for object localization problems in an automated production line,we designed a embedded machine vision system with an ARM(Advanced RISC Machine)and FPGA(Field-Programmable Gate Array)embedded processor,and propose a FPGA-based connected components analysis algorithm using non-equivalence tables,that can improve the running speed of its image processing.The main contents of this thesis are summarized as follows:(1)We designed an embedded machine vision scheme based on FPGA with a processor core,and completed the hardware platform of embedded machine vision system.What's more,the system has fulfilled the basic functions including image acquisition,ARM-FPGA data interaction and data output,the separate experiment shows that the embedded machine vision system which has the basic functions can run correctly.(2)Combined with the parallel computing performances of FPGA and the existing analysis algorithm of connected components,we proposed a FPGA-based connected components analysis algorithm using non-equivalence tables.Through the validation experiment,it figures out that our algorithm has the same image analysis results with PCSoftware method but faster processing speed.(3)The proposed algorithm is integrated with the embedded machine vision system which has the basic functions,furthermore the target localization experiments for the whole system was designed and analysed.Results show that the processing period is 127 ms per single image with the size of 512*512.(4)The whole system is applied into the Delta parallel robotic sorting systems,and the results confirmed the practical value of the system by completing the task of target localization.