Burr Contour Recognition and Coordinates Sequence Real⁃Time Generation for Robotic Deburring Efficiency Optimization

The phenomenon of burring is common in the manufacturing of metal parts.This phenomenon directly influences the assembly accuracy and service performance of the mechanical parts.In this work,we propose a vision⁃based method for two⁃dimensional planar workpiece.The proposed technique has the ability to recognize burr contour and generate the coordinate sequence in real⁃time along x and y directions.The robotic deburring efficiency is improved based on the quantitative information of the burr size.First,by utilizing the local deformable template matching algorithm,we match the standard workpiece contour with the workpiece contour to be processed and compute the corresponding pixels distance between the two contours.Second,we set the distance thresholds in order to divide the burr contours into different levels.We extract the coordinates of the burr contours and map them to the standard workpiece contour.As a result,the closed⁃loop robotic deburring path sequence is generated.Finally,on the basis of the quantitative information of burr size,we adjust the deburring speed in real⁃time during the deburring process.The experiments performed in this work show that the deburring time of the proposed method is reduced by 15.45%,as compared with the conventional off⁃line programming deburring methods.Therefore,for industrial mass production,the deburring efficiency is greatly improved.

A CAD-BEM geometry transformation method for face-based primary geometric input based on closed contour recognition

Performance analysis during the early design stage can significantly reduce building energy consumption.However,it is difficult to transform computer-aided design(CAD)models into building energy models(BEM)to optimize building performance.The model structures for CAD and BEM are divergent.In this study,geometry transformation methods was implemented in BES tools for the early design stage,including auto space generation(ASG)method based on closed contour recognition(CCR)and space boundary topology calculation method.The program is developed based on modeling tools SketchUp to support the CAD format(like*.stl,*.dwg,*.ifc,etc.).It transforms face-based geometric information into a zone-based tree structure model that meets the geometric requirements of a single-zone BES combined with the other thermal parameter inputs of the elements.In addition,this study provided a space topology calculation method based on a single-zone BEM output.The program was developed based on the SketchUp modeling tool to support additional CAD formats(such as*.stl,*.dwg,*.ifc),which can then be imported and transformed into*.obj.Compared to current methods mostly focused on BIM-BEM transformation,this method can ensure more modeling flexibility.The method was integrated into a performance analysis tool termed MOOSAS and compared with the current version of the transformation program.They were tested on a dataset comprising 36 conceptual models without partitions and six real cases with detailed partitions.It ensures a transformation rate of two times in any bad model condition and costs only 1/5 of the time required to calculate each room compared to the previous version.

An Algorithm to Recognize the Target Object Contour Based on 2D Point Clouds by Laser-CCD-Scanning

For a vision measurement system consisted of laser-CCD scanning sensors, an algorithm is proposed to extract and recognize the target object contour. Firstly, the two-dimensional（2D） point cloud that is output by the integrated laser sensor is transformed into a binary image. Secondly, the potential target object contours are segmented and extracted based on the connected domain labeling and adaptive corner detection. Then, the target object contour is recognized by improved Hu invariant moments and BP neural network classifier. Finally, we extract the point data of the target object contour through the reverse transformation from a binary image to a 2D point cloud. The experimental results show that the average recognition rate is 98.5% and the average recognition time is 0.18 s per frame. This algorithm realizes the real-time tracking of the target object in the complex background and the condition of multi-moving objects.