Research On Recognition Method Of Assembly Operations Based On Two Dimensional Sensor

Human-Computer Interaction has become a major issue that urgently needs to be resolved in the information field in the 21st century and a focus of current competition in the information industry.Countries around the world have regard human-computer interaction technology as a key technology for key research.This subject is oriented to mechanical assembly operations,taking the assembly of the reducer shaft system as the research object,launching the research work on the recognition of the operation gestures in the assembly operation,analyzing the changes of the hands in the assembly operation,and combining the complex and changeable,occluded,and mixed assembly operation gestures.Features such as overlapping and similar postures,two-dimensional collection of gesture data for assembly operations,determination of a reasonable two-dimensional recognition scheme,fusion of gesture data,extraction of gesture features,and completion of gesture recognition with the help of key frame extraction algorithms and recognition algorithms,fast Accurately distinguish the assembly operation action,strengthen the ability of teaching and human-computer interaction.The main research work of the thesis is as follows:(1)Aiming at the problems of low recognition accuracy and slow speed in assembly gesture recognition,an action recognition solution based on two-dimensional data fusion is proposed,and the principles and key technologies involved in the solution are analyzed.By decomposing assembly tasks,the relationship between hand motions,motion elements,and assembly operations is established,and an action recognition technology route is formed with the division and extraction of work gestures,two-dimensional gesture data collection and fusion,and work gesture classification and recognition as the core.And analyze the key technologies such as multi-dimensional data fusion and key frame extraction algorithm to ensure that the technical solution is feasible and help to improve the accuracy and speed of gesture recognition.(2)The thesis takes the assembly of the shaft system in the reducer as an example,and analyzes the assembly operation process.Using the analysis of the motion element,the 12 assembly operation gestures involved in the shaft assembly are extracted,and they are divided into static gestures according to the displacement changes,and dynamic gestures;Gesture modeling for human hand structure analysis,completed the extraction of gesture features,used to characterize assembly operation gestures;Finally,for the distortion of collected data due to occlusion,aliasing and other phenomena,a dual Leap Motion layout design was carried out.A two-dimensional data acquisition system was built to collect and process the gesture data of assembly operations.(3)A two-dimensional-based assembly operation action recognition method is studied.Select the SVD coordinate system conversion method to perform coordinate conversion on the gesture data.For the different placement angles of the dual Leap motion,the coordinate rotation matrix and translation matrix at different angles are obtained using standard gestures;the coordinate deviation is performed on the gesture data obtained under the same coordinates after the conversion.Through analysis,the optimal placement angle of dual Leap motion is determined;the gesture data fusion method based on principal component analysis is studied,and SVM is used to train and learn the fusion gesture features.Through the comparison and analysis of recognition results,it is proved that data fusion can improve gestures.The accuracy and reliability of recognition.(4)The method of gesture recognition for assembly operations based on key frames is studied.The comprehensive change characterization parameters of each frame of gesture data are designed,combined with the time series to form the change curve of the operation action;the local extreme points of the curve are searched,and then density clustering is performed to obtain the key frames of the gesture sequence;in determining the key frames Case studies were carried out on the basis of the extraction algorithm process,which realized the key frame extraction and recognition of the gesture motion sequence of the bearing assembly operation,completed the analysis of the matching degree between the recognition result and the standard operation sequence,determined the credibility of the recognition result,and proved the feasibility and effectiveness of the method.