光学
激光器
惯性约束聚变
校准
材料科学
镜面反射
融合
物理
语言学
量子力学
哲学
作者
Huijun An,Lingbao Kong,Min Xu,Huixin Song,Xintao Xu
标识
DOI:10.1016/j.optcom.2023.129627
摘要
Line laser is widely used in structural light three-dimensional (3D) measurement of metal parts; however, the laser stripes are severely disturbed when measuring complex surfaces. In the process of laser plane calibration, the laser stripe will be damaged by the pattern on the calibration board; when measuring metal surfaces, the laser stripe will be exposed unevenly. Overexposed parts show highlight and artifact, while underexposed stripes appear burrs, high noise, or even undetectable due to material absorption, which will greatly affect the accuracy of stripe center extraction. In this paper, a modified multi-exposure fusion (MMEF) method is proposed to improve the dynamic range of laser stripe images by enhancing underexposed regions and reducing saturation. The MMEF method uses a “smoothness” index and a “validity” index to give more weight to smooth, Gaussian-like laser stripe patterns and extract useful modulation information. To evaluate the effectiveness of our method, the performance of the MMEF method is tested for laser plane calibration and metal surface reconstruction, and the high dynamic range (HDR) method, retrieval fusion (RF) method and exposure fusion (EF) method are introduced for comparison. Measurement experimental results show that the MMEF method proposed in this paper improves the laser plane calibration accuracy with 86.0% and the 3D measurement accuracy of metal surface with 12.5%.
科研通智能强力驱动
Strongly Powered by AbleSci AI