Research On Vision-based 3D Measurement System For Large Ship Engine Forgings

The engine is an important core component of the ship,and the quality and manufacturing level of its forgings determine the reliability and life of the engine.Efficient and accurate three-dimensional digital measurement of the shape and size of forgings is of great significance for improving the forging quality and processing efficiency of forgings.Therefore,a line structured light vision three-dimensional measurement experimental platform to measure the shape of forgings were designed and built.Appearance and dimensions are measured.In the online structured light vision measurement of large ship engine forgings,the light bar center extraction is the key technology,which directly affects the measurement accuracy.An improved gray-scale gravity center extraction algorithm was used to reconstruct the shape of the forging after system calibration,and measured the key dimensions of the forging.Experiments show that the accuracy of the improved gray-scale gravity center extraction algorithm has been greatly improved,and the measurement accuracy has also been improved.The specific work was as follows:(1)In the online structured light vision measurement,due to the influence of complex conditions such as the quality of the line laser and the surface roughness of the forging,the poor extraction effect of the center of the light strip leads to inaccurate 3D measurement results.Therefore,an extraction algorithm based on improved gray barycentric method was proposed.Firstly,the image was preprocessed,and the light stripe skeleton was extracted as the initial center point.Then the direction template method was used to determine the normal direction of the initial center point.Finally,the weighted gray barycentric method with adaptive width was adopted to perform the initial center point of the normal position.The sub-pixel coordinates of the center of the light strip projected on the forging surface are obtained.Experiments show that the accuracy and efficiency of the improved algorithm were improved compared with the traditional center extraction algorithm.(2)The calibration of camera,line structured light and hand-eye relationship was studied and completed.Based on Zhang’s comprehensive multi-distortion iterative calibration method,the internal and external parameters and distortion coefficients of the camera were obtained.A line structure cursor positioning method based on the camera perspective model was established,which can improve the accuracy of fitting the light plane by improving the accuracy of feature point extraction and increasing the number of feature points.Experiments show that the accuracy and robustness of this method have been greatly improved.By establishing a hand-eye model based on a single-point constraint,the hand-eye calibration of the robotic arm was achieved.After system calibration,various parameters in the measurement model were obtained.(3)The preprocessing methods such as noise reduction and sparseness of multi-view3 D point clouds were studied,and the splicing of point clouds under different fields of view was realized by using marker-based splicing technology combined with the least square method.Based on vs2015 + QT,the 3D vision scanning measurement system was designed and developed,and the software integration of each function was realized.By pasting the marking points on the surface of the workpiece,the forgings under different viewing angles were scanned,and the acquired point cloud was preprocessed.Then the point clouds on the surface of the forgings were spliced under different viewing angles.And the whole point cloud model of the forgings was obtained.Then the three-dimensional measurement experiment of the forgings was completed.The results show that the 3D point cloud and the key dimension information of forging can be obtained by the system.