Research On Inspection And Positioning Of Bearing Cap Based On 3D Vision

The "Made in China 2025" strategy raises the requirements for the national manufacturing industry's independent innovation capabilities and accelerates the industrial upgrading of manufacturing industries such as automobile production and processing.The crankshaft bearing cap is one of the key parts of automobile engine.The precise combination of the cover and the cylinder block is the prerequisite for the high-speed and stable operation of the crankshaft.Therefore,pressing the bearing cap on the cylinder block has become an important step in the production process of automobile engines.At present,the detection and installation of the crankshaft cover still adopts manual methods in China,which have problems of strong subjectivity,inconsistent standards,high labor costs,and high labor intensity.Intelligent systems and methods are required to replace the traditional methods for automatic installation and detection of bearing caps.Aiming at the need for autonomous operation of the bearing cap assembly system,this thesis proposes a method for detecting and positioning the bearing cap group based on 3D vision to assist the robot in loading bearing caps.The main research contents of this article are as follows:1.According to the production environment and requirements of the automobile engine plant,and the material and structural characteristics of the crankshaft bearing cap,a corresponding lighting imaging scheme and supporting device structure are designed for the automatic take-up robot of the bearing cap.The system can effectively collect and detect images,and control the robot and mechanical and electrical modules through signals,which cooperate to complete the automatic loading of the workpiece of the bearing cap group,and realize the automatic detection and loading of the crankshaft bearing caps.2.Aiming at the difficulty in detecting the orientation and the order of bearing cap groups,an online detection method for the bearing cap group based on fuzzy inference system is proposed.Firstly,pre-processing such as channel conversion and median filtering is performed on the image on the roof surface to suppress noise and highlight features.Secondly the edge template is extracted by Canny operator.Finally The bearing cap group is identified by matching the template of the image pyramid,and then the positions of the bolt holes and ink lines on the bearing cap are calculated.Based on this,the input / output variables,membership function and fuzzy rules are designed based on fuzzy theory,and the method of fuzzy decision is used to determine the direction and order of the bearing cap group.3.A 3D visual positioning method based on line laser is proposed.Based on the two-dimensional template matching,the secondary laser imaging is performed by a linear laser that is angled to the camera,and then an improved Steger algorithm is proposed to extract the centerline of laser light bar.Finally,the laser image is synthesized,then the posture and height information of the bearing cap group is calculated.The calibrated transformation matrix can determine the position and posture of the bearing cap group in the world coordinate system,so as to guide the picking robot to grasp and place the workpieces.Test experiments prove the real-time and effectiveness of the 3D visual positioning method proposed in this thesis.