Width Measurement and Head Tail Shape Detection for Iron Slab using Machine Vision

In the production of steel, lateral bending, also known as Camber, is a major defect. However, to achieve production of desired and high quality steel, the ability to accurately measure the strip and plate width, detect the head tail shape of strip slabs and to remove lateral bending during hot rolling is of essence. This is also paramount in achieving ace outturn. In this study, head tail shape detection and width measurement of slabs using machine vision and image processing techniques is developed. A model was developed to measure the width, head tail shape of an iron slab using machine vision. Firstly, impulse noise was removed by computing local similarity analysis and calculating the neighborhood noise evaluation function Gaussian noise is filtered out by combining the inverse of the gray difference together with the reciprocal of the geometric distance. After which improved canny edge detection is implemented. Midpoint and width is measured estimated by using progressive scanline based on edge points. Lastly, the head tail shape is calculated using the measure of depression on the steel plate along the length of the steel direction. The program computes the distance between the most depressive point and straight line that links the two end of the center line. Four simulated slabs are used to test the proposed method, experimental results proved the effectiveness of the study.