Research On Dimension Measurement And Shape Detection Technology For Large Forging

As the key and core basic component, large forgings are required for national heavy manufacturing equipments and major construction projects, which directly determine the manufacturing and development of major technology equipments. In the process of forging and machining, real-time and accurate online measurement and estimation of heavy forging are of great significance for improving quality and efficiency and reducing material machining allowance in both theory and practice. According to the current situations of backward measurement techniques, great errors and severe measurement environment for heavy forging in our country, the paper gives a detailed research on dimension measurement and shape detection technology. Based on the binocular vision measurement system of multi-line structured light, this paper is to explore new ideas and methods for non-contact measurement of large forging.First of all, combining with domestic and overseas research status of dimensional measurement of heavy forging, this paper analyzes a great amount of literatures related to non-contact measurement techniques. Aiming at the characteristics and existing problems of current major heavy hot forging, the method of binocular vision measurement based on multi-line structured light is proposed to measure large forging by the way of non-contact measurement, which is more easily to obtain information about large forging measurement.Secondly, according to the spectrum selective image acquisition technology, an experiment platform for image collection of large forging is designed and built. In order to filter self-emitted radiation of hot parts, the strong radiation and interfered lines should be filtered by placing short-pass filter in front of lens of two-CCD camera, the original images and structured light stripes are retained. The forging images of multi-line structured stripes which are close to room temperature can be acquired.Thirdly, the paper uses Canny edge detection operator to extract the edge of light stripes and acquire edge with higher precision. Also, the paper uses eight-pixel neighborhood algorithm to extract image pixel coordinates from edge. On the basis of the polar geometry constraint theory, the paper analyzes the deducing process of epipolar equation and fundamental matrix and puts forward the matching algorithm of stripes edge under the principle of epipolar constraint matching criterion. It realizes corresponding points matching for the pixel points of left and right light stripes edge and achieves a relatively good result with image matching points. Based on the three-dimensional reconstruction mathematical model of point cloud, point cloud of the rectangle and cylindrical workpiece under experiment conditions is reconstructed, which can accurately reflect the shape and physical dimension of forging.Finally, according to the problem of cylindrical forging diameter measurement, the paper proposes that the least-squares ellipse fit methods should be applied to curve-fitting of stripes points of cylindrical workpiece. The acquired point cloud of cylindrical forging is transformed from three-dimensional space to two-dimensional plane through plane fitting and rotation transformation in space. Also, the method of least-squares ellipse fitting is used for fitting stripe point cloud to work out related ellipse curve equation and obtain the diameter parameter. In the laboratory, it carries out simulation fitting measurement. The experiment result demonstrates that the method in this paper meets precision and requirements of heavy forging manufacturing industry with small errors, which realizes the diameter measurement of cylindrical forging.