A 3D Measurement Method with Accurate Boundaries Based on Mutation Feature Detection of Laser Stripes
Yang Zhang,Wei Liu,Yongkang Lu,Hongtu Di,Xikang Cheng,Sitong Liu,Peng Gao,Fuji Wang,Zhenyuan Jia
DOI: https://doi.org/10.1088/1361-6501/aafd85
IF: 2.398
2019-01-01
Measurement Science and Technology
Abstract:Laser scanning measurement plays an important role in detecting the parameters of parts in a workshop. However, the existing methods and measurement instruments are unable to accurately calculate the 3D coordinates of large-scale part boundaries in an on-site complex environment. This paper proposes a 3D measurement method with accurate boundaries based on the mutation feature detection of laser stripes. First, drawing on the type analysis of light stripe images in the target boundary, the corresponding ideal boundary models are built. Next, an adaptive classification mechanism is designed in order to achieve rapid recognition and classification of the mutation features of laser stripes. Then, a boundary endpoint laser stripe detection method based on mutation features is presented as a way to obtain sub-pixel boundary endpoint locations. Finally, simulations and lab experiments are conducted using a standard part to validate the accuracy of the proposed method. In the analysis of results between the measured data and the reference distance, the proposed method demonstrates a high measuring accuracy, with a root mean square of the relative error of 0.024%. This measuring accuracy satisfies the requirements of field measurement. In addition, a measurement experiment with a curved surface part is carried out, which further verifies the feasibility and effectiveness of the proposed method.