Research On Digital Image Identification To The Bend Distortion Of Shaft Forgings

Since bend distortion often occurs in forging process of shaft, especially of those having axial dimension much longer than the radial. a rectification procedure is often used in oder to reduce the costs of materials and man-hour of machining. It is very necessary to measure the pattern and parameters of the bending in automatic rectification and product inspection. Because he manual measurements being used are of difficult conduction and low efficiency, it is significant to implement long distance. non-contact measurement. Based on the investigation into the measurement of heavy forgings in the world, a CCD based, image measurement approach of forging bending is researched.The whole research include hardware equipment design for measurement system, algorithm research for important problem, data processing, software program design, device calibration and mesurement expriment. In this thesis, adjustment for uneven illumination, image filering and technique for image enhancement are analyzed; algorithms for image segmetation are researched and a mathematical morphology method based on threshold segmetation to reduce aberration is put forward after the common edge detection algorithm is not satisfied, the results show that the method achieved a better segmentation results, image segmentation and forgings border; a mathematical model for calculating bending parameters of forgings, based on single-vision, is established and bending parameters are obtained from centerline three-dimensional reconstruction.In the experiment, the measurement system is calibrated with a easy and practical method and a three-dimensional coordinate measuring machine is used for verifying the measurement results of bending. The results show the agreement of their data and satisfied achivements.Research for the methods of forging bending measurement in this thesisprovides a base foundation for futher research and application to practic.