Research On Contour Precision Detection Technology For Complex Sheet Parts Based On Coarse-fine Progressive Machine Vision

Thin-sheet parts are a class of two-dimensional parts with well-defined contours and small thicknesses,which are widely used in precision machinery,electronic instruments,3C products and other fields.The contour shape of thin-sheet parts directly affects its function.Due to the small stiffness,thin-sheet parts are prone to deformation in the manufacturing process.In order to ensure product quality,it is necessary to detect its contour shape.Machine vision detection has the advantages of high efficiency and good flexibility.It is widely used in the field of part shape detection.However,its measurement accuracy and measurement range restrict each other.It is difficult to ensure the measurement accuracy for the measurement object with relatively large contour size.Therefore,it is of great significance to study a large-range,high-precision and highefficiency contour detection method for thin-sheet parts to improve the detection level of parts and the quality of product manufacturing.In this paper,on the basis of CMM,taking complex thin-sheet parts as the object,the coarse-fine progressive machine vision method is used to carry out a more in-depth study on the precise detection of the contour of the thin-sheet parts.Aiming at the problem of mutual restriction between measurement accuracy and measurement range in vision measurement,a coarse-fine progressive vision measurement method was proposed,and a multi-lens integration device was designed for the physical integration of different magnification lenses.Firstly,the low magnification lens is used for rough measurement,and the panoramic image of the measured thin-sheet parts is collected,and the contour data of the parts are extracted by image processing technology.Then the fine measurement path planning is carried out according to the extracted contour data.Next,the local tracking measurement is carried out according to the fine measurement path by the high magnification lens to obtain the local image of the part contour.Finally,the high-resolution panoramic image of the thin-sheet part is obtained by image stitching,and the contour parameters are analyzed.Aiming at the problem of fine measurement path planning,a fine measurement path pre-planning method based on "tiling intersection method" is studied,and an adaptive genetic simulated annealing algorithm is proposed to optimize the path.Firstly,the coordinate transformation relationship between coarse and fine measurement is constructed to transform the field of view of high and low magnification lenses.Then the center coordinates of the field of view of the local tracking measurement are calculated based on the ’ tiling intersection method ’ to obtain the pre-planned path.Finally,the adaptive genetic simulated annealing algorithm is used to iteratively optimize the measurement path to obtain the optimal measurement path.Aiming at the problem of local image stitching of thin-sheet parts,the template selfsearch method is studied,and the automatic stitching of local images is realized by gray template matching.First,by customizing the template size,traverse the local image to obtain an image template that intersects with the part contour.Then the similarity between each template and the image to be matched is calculated by NCC algorithm.Finally,the template with the largest similarity value is taken for image stitching.Based on the theoretical research,a coarse-fine progressive machine vision detection experimental system was constructed,and the experimental verification was carried out with the sheet gear as the measurement object.The experimental results show that the proposed coarse-fine progressive machine vision measurement method can achieve high efficiency and high precision measurement of the contour of complex thin-sheet parts.