Research And Implementation Of The Algorithm Of The Accurate Measurement Of Billet Baseed On Dynamic Scenes

Steel is an important raw material in the production processes, the level of a country’s iron and steel industry has a serious impact on the construction of the national economy. With the rapid development of the national economy, China’s demand for steel is growing. Continuous casting system in steel production process is the core of the steel industry and is an important indicator to measure the level of steel production. Domestic and foreign iron and steel industry regard the development of the continuous casting system as a top priority of the production of iron and steel industry. In the steel industry, which is a very important part from the ladle out of billet to be cut, this cut is a fixed-length line cut. Its cutting accuracy of a direct impact on the finished product rate of the billet.Then the most important step in the cutting process is the measurement of its length.With the rapid development of computer technology, computer vision-based measure-ment technique has been widely used. This paper will use the digital image processing technology to measure the target in a lab environment. In this paper, image filtering, system calibration, the edge of the fitting and the edge of the corner extraction technology are in-depth analysis and research, and on this basis for this paper possible ways.In the industrial measurement, the influence of lens distortion brings measure a great deal of measurement error, so the image distortion must be corrected before conducting accurate measurements. In teams of image correction.First, the distortion coefficient of the camera is calibrated through a template, and then according to their coefficient of the image distortion correction.In terms of image filtering, we analyze the source of the noise. According to characteristics of noise, we use median filtering techniques and eliminate the interference of noise while maintaining the edge of the image characteristics. In the calibration, the calibration long point calibration of the system.In this article, the first extraction of calibration points, the center of the calibration points as standard of known length of the calibration system. Edge extraction using a least squares fit, especially the extraction of the edge, the first fitting a straight line, find the edge points to the square of the deviation of each straight line, select the smallest piece of the deviation of a straight line as the boundary. Minimize the error caused due to boundary extraction. Extraction in the corner, corner features template based on linear characteristic features of the target binary image, according to template to extraction target corner. Finally establishment of an offset correction model, the objectives of the offset calibration line length correction. End of the article gives the experimental results measured in the laboratory environment, according to measurements the different locations of the standard analyte; the results basically achieve the expected effect of measurement.