Research On Cold Rolling Strip Shape Detection Method Based On Image Processing

As the most basic raw material,steel plays an important role in the development of our national economy and society.Various industries not only for steel and other metal raw materials demand increases sharply,the demand for steel precision is more and more high,especially for the horizontal and longitudinal thickness of the plate precision requirements increase.For cold rolling with the demand of the industrial field of flatness detection and control,put forward a kind of flatness detection scheme based on image sensor,in an appropriate location to install and image sensor,laser line laser can make from the laser irradiation to the steel band,captured by image sensor can reflex the light strip and line of the laser.Finally,the shape information is obtained by image processing.In this paper,the principle and method of many common flatness detection schemes are analyzed,and their advantages and disadvantages are compared.Combined with the actual situation of the field,the non-contact shape detection scheme based on image processing is selected.In the process of image preprocessing,color space conversion,region of interest extraction,bilateral Gaussian filtering,image enhancement and other methods are used to preprocess the image.A variety of image edge detection algorithms were compared.On the basis of giving priority to the performance of the algorithm,Canny algorithm was selected to detect the edge position of the laser line,which was used as the reference range of the center position detection of the laser line structured light.At the same time,the width of the laser line was calculated.In this paper,the traditional gray gravity center method is improved.Finally,the gray gravity center method based on edge detection results is used to calculate the laser line center position.Hough line detection algorithm is used to accurately detect the edge position of strip steel.The detection range of strip steel shape is reduced and the accuracy of strip steel shape detection is improved.Finally,the strip shape information is calculated by the center position of the laser line.The real-time strip shape state is restored.The experimental results show that the strip shape information can be obtained by using this method,which lays a foundation for the further accurate control of strip shape.