Centerline Extraction Method of Variable Width Light Strip Based on Improved Gaussian Fitting
Line structured light is one of the most widely used techniques in 3D measurement,and its measurement accuracy is mainly determined by the extraction accuracy of the Central Line of the laser strip. However,in practical applications,the height change of complex object surface will cause the width of the light strip to change,resulting in the gray distribution of the light strip section no longer meets the standard Gaussian distribution. The traditional centerline extraction algorithm is often lack of robustness in the face of such non-uniform width and non-standard light strips,resulting in the decline of accuracy of centerline extraction results. In this paper,a new Gaussian iterative fitting algorithm is proposed to improve the stability of centerline extraction and thus improve the measurement accuracy. Firstly,the light stripe region is segmented and the initial width of the light stripe is estimated,and the gray distribution characteristics of the light stripe are improved by Gaussian filtering. Then,the gray level of the light stripe is iteratively fitted by section,which includes the parameter initial value optimization process to eliminate the influence of the variable width of the light stripe. The results show that the improved Gaussian fitting algorithm proposed in this paper has high stability and accuracy,and achieves higher accuracy than the traditional method in the noise environment,which can effectively improve the accuracy of the line structured light measurement results.
centerline extraction of light stripGaussian fittingvariable width light stripiterative algorithm
万方数据
维普
