Study On Flatness Measurement Method Based On Onboard FPGA System For Hot-rolled Strips

Strip steel is one of the de-capacity products in China's steel industry not only because of the lack of reasonable prediction for the capacity in the market and timely adjustment of the structure of steel production,but also mainly since our strip steel is not of high-quality and the innovation of production process is not enough.Therefore,producing high-quality strip steel and promoting the steel industry from large to strong have been important measures in our supply-side structural reform.In recent years,hot-rolled strip technology is universal in the market.As an important quality indicator in strip steel production,flatness measurement and its detection technology attract much attention and develop for a real-time high precision detection.In order to obtain the hot-rolled flatness measurement's precise detection,this paper has carried on the following research and obtained the corresponding results based on the graphical FPGA programming technology.Firstly,through the analysis of the relationship between the laser source and the camera position,consider the effect occlusion and shadow have on detection precision;adopting the direct laser range of triangle,the detection of strip steel flatness measurement is switched to the height measurement at any point on its surface,which builds the corresponding relation between the coordinate points in world coordinate system and image plane pixel coordinate system;Then appropriate distortion model can be chose according to the actual situation experiment contrast and through the calibration algorithm of camera,the internal and external parameters and distortion parameters of cameras can be get for the subsequent system detection,which can correct lens distortion effectively.Secondly,based on the onboard FPGA image acquisition card,using Visual Applets graphical FPGA and Visual Studio 2010 C++ programming software establish the flatness measurement on-line detection system.Through simulation analysis and theoretical calculation,the hardware is verified for the feasibility and superiority.At the same time,algorithm design and its implementation should be focused on using FPGA to operate the underlying hardware for a programming simplification and system architecture.Thirdly,a detection method based on improved gray weighted centroid laser stripe is put forward.The graphical programming realizes the corresponding algorithm and then can directly generate SDK files,which will be included in the C program;using API function can obtain the coordinate information of the laser stripe,so we can analyze the flow of image information in real time and understand every process of image processing.At last,since laboratory already has detection model which applied to industry production and related research results,the introduction of FPGA card in the actual detection system can greatly enhance its timeliness and accuracy.