Font Size: a A A

Research On Automatic Detection And Assembly Control Software Of Miniature Parts

Posted on:2019-11-17Degree:MasterType:Thesis
Country:ChinaCandidate:W WangFull Text:PDF
GTID:2382330563958748Subject:Mechanical engineering
Abstract/Summary:PDF Full Text Request
Micro inertial accelerometers are widely used in aerospace,autopilot and other fields because of its excellent performance,stability,reliability,low power consumption and other advantages.However,the inertial accelerometers with complex structures,whether assembled by hand or by man-machine interactive device,all are in low productivity.In this situation,an automatic assembly strategy is designed based on the precision assembly equipment to realize the automatic assembly of parts.The assembly task is to realize the accelerometer final assembly.The automatic assembly process of the parts is designed according to the assembly accuracy requirements.The assembly system is designed based on the modular idea,including a vision measurement module,an assembly operation module,a thread assembly module,a working table module,and a feeding module.It adopts the visual and guided mode of Look and Move.The vision module measures the position and orientation information of parts and guide the other modules to complete the assembly work.Visual inspection is the key to automatic assembly.Because of the short depth of field,the vision system is prone to defocus.An autofocus scheme with a coarse-fine combination strategy is designed to solve the problem.The Variance function and the frequency function are selected as the sharpness evaluation function of the image.The hill-climbing method and two times interpolation method are used as a focus search strategy.The autofocus scheme can ensure the image quality that the visual system collects.For a collected part image,a reasonable processing method is designed to get the assembly information.Gaussian filtering and median filtering are used to smooth the image and reduce the noise.The multi-threshold segmentation based on particle swarm optimization and Canny edge detection are used to distinguish the foreground and background of the image.Hough transform and blob detection are used to identify image target and detect part position.According to the assembly process,the assembly program is developed.The program is designed with a layered architecture,which uses the feedback information and pre-calibrated parameters to determine the execution status of each assembly step without manual intervention.According to the modular design concept,codes for a type of function or for a specific hardware are packaged to reduce the difficulty of developing program and make the codes reusable.A series of parts were used to perform assembly experiments on the automatic precision assembly system.The results showed that the assembly efficiency is greatly improved compared to the manual assembly.
Keywords/Search Tags:Miniature Parts, Autofocus, Image recognition, Automatic assembly
PDF Full Text Request
Related items