Research And Development Of Image-based Static Leveling System

In this paper,the image processing technology is introduced into the static level measurement system,and the industrial camera is used to read the buoy position in the crucible to determine the static level coordinates,and the static level measurement is completed.The author made the following results by making the physical measurement model and using Matlab software to do the buoy image preprocessing,the buoy circle detection algorithm simulation,the liquid level measurement mathematical geometry model analysis and the measurement accuracy calibration experiment.(1)The study found that the use of dark and lightweight spherical buoys can make the liquid level image analysis identification and positioning more accurate,and the surface of the buoy should be covered with hydrophobic materials to reduce the measurement error caused by liquid adhesion;the viscosity of distilled water is relatively low,suitable for The liquid used in the static level measurement in this subject;the backlight field illumination is needed when collecting the buoy image;compared with the ordinary civilian camera and CCD camera,the COMS camera is more suitable for the buoy image acquisition of this article.(2)The pre-processing experiment is carried out by buoy images simulating the measurement.The results show that the buoy image collected by the industrial camera in the image static leveling system mainly has Gaussian noise and Salt and Pepper noise.It needs to use median filtering and mean filtering to do denoising processing.In addition,image preprocessing work such as linear gradation transformation,OTSU threshold segmentation,morphological operations and Canny operator edge detection are also needed.(3)The simulation results are used to analyze the advantages and disadvantages of the least squares fitting circle and the traditional Hough transform detection circle,and the traditional Hough transform detection circle algorithm is optimized and improved by the least square method and the expansion circle method.The simulation results show that the improved algorithm has excellent robustness when detecting images with pseudo edges and noise points,and the detection accuracy reaches 0.05 pixels.(4)Based on the small hole imaging model,the distortion of the industrial camera is reduced by the camera calibration experiment,and the liquid level measurement resolution of two different shooting modes is analyzed based on the principle of small hole imaging.The simulation results show that the camera can obtain better liquid level measurement resolution by shooting perpendicular to the buoy motion trajectory.(5)In this paper,the liquid level measurement accuracy experiment and the static level measurement accuracy experiment are carried out.The projection surface offset correction model is established and the absolute coordinate correction formula is summarized to correct the measurement results.The final experimental result data shows that when the ranges are 0-30 mm and 0-20 mm,the corrected measurement accuracy of both experiments reached 0.5mm,and the medium error of the measurement experiment was 0.25 mm,it meets the accuracy standard of second-class static level measurement.