Linear Displacement Measurement Model Correction And Displacement Solution Based On Image Grayscale Information

Displacement measurement techniques cover many areas of modern science.In recent years,the development of ultra-precision manufacturing processing has placed higher demands on the speed and accuracy of displacement measurement.With the advancement of digital image processing technology,precision displacement measurement based on image sensor has attracted the attention of scholars of domestic and overseas due to its non-contact,full-field measurement and high precision.In this paper,linear displacement measurement is taken as the research object,and a measurement method for solving linear displacement by redundant detection information of multiple image sensors is researched.The influence of the n?mber of sensor on the measurement accuracy is analyzed in the simulation,and the performance of the measurement method under noise is analyzed.The filtering algorithm of the detection signal is researched and verified in the experiment.In the actual measurement,the distortion of the detection model relative to the ideal model leads to the reduction of the accuracy of the model identification,which finally reduces the measurement accuracy.The unevenness of ill?mination and reflection,the geometric distortion of the imaging system,and the intensity of the image sensorgrayscale value quantization are analyzed.A model parameter correction method based on high-order polynomial is proposed,the method improves the model accuracy by reconstructing the distortion function superimposed on the ideal model.The experimental results indicate that the method significantly improves the accuracy and measurement accuracy of the model.Aiming at the inconsistent response of each sensor in the actual measurement of sensor array,a correction method based on two-dimensional polynomial is proposed.The inconsistency of each sensor is corrected while correcting the influence of the above three kinds of distortion factors.The experimental results demonstrate the effectiveness of the method and further improve the measurement accuracy.In the actual measurement process,due to the installation factor,the sampling direction of the sensor array cannot be completely parallel with the moving direction,so there is an angular error.Inconsistent sensor spacing results in pitch errors.Aiming at the above problems,the influence of them on the model parameters is analyzed.A dynamic coefficient correction method combining Fourier series and high order polynomials is proposed.The experimental results indicate that the proposed method can better correct the various model distortion factors mentioned above,and the model accuracy is further improved.Under the measurement stroke of 10.46 mm,the standard deviation of measurement error achieved to 1.5?m.Aiming at the different effects of the detection information of each sensor in the measurement process on the displacement solution,a method called multi-pixel weighting is proposed,in which each pixel independently estimates the displacement,and the estimated values are weighted and combined to obtain the final estimated value.The method dynamically solves the optimal weight of each estimated value during the measurement process,so that the detection information of each pixel can be reasonably utilized,and the optimality of the displacement estimation value at each moment is satisfied under the criterion of the least square error.The experimental results show that the standard deviation of the measurement error of the method under the measurement stroke of 8.6 mm is 1.5 ?m.