Sub-Pixel Centerline Extraction Based on Light Stripe Adaptive Grayscale Weighting

Gailing HU,Xiang ZHOU,Tao YANG,Junping WANG,Chunwei ZHANG,Gangfeng WANG,Yuqin LI,Jiayu GUO

DOI: https://doi.org/10.7652/xjtuxb201805003

2018-01-01

Abstract:This paper presents a method of adaptive grayscale weighted center extraction of light stripe based on skeleton tracking in order to improve the linear structured light three-dimensional surface measurement accuracy.First,a set of fairing and unidirectional skeletons is calculated by mathematical morphological method from the acquired images.Taking the skeleton as a search path,a useful pixel region is set along the normal direction of the skeleton.The grayscale threshold is divided according to the degree of importance of pixel grayscale on the light stripe centerline extraction in the normal section of the skeleton.The value of pixel gray which is more important to centerline extraction is weighted square.Then,the sub-pixel center of the light stripe is calculated according to the gray weighted centroid method.Finally,simulations and experiments are carried out.The simulation results show that the average deviation from the centerline of the light tripe to the fitted curve is less than 0.3 pixels,the standard deviation is less than 0.03 pixels,and the calculation time is 87.6 ms.The experimental results show that the maximum deviation,average deviation and standard deviation extracted by this method are all smaller compared with other methods.The proposed method has some advantages,including high accuracy,stable extraction result and good robustness,whereby a sub-pixel light stripe centerline with fine fairness can be obtained and hence the online measurement requirements could be satisfied.