Code-based fringe projection profilometry (CBFPP) is widely used because of its simple structure, robustness, and potential high speed. One of the most important steps in this technique is to obtain the fringe orders correctly. However, fringe orders tend to fail easily, especially around the phase-jump regions. First, this paper analyzes the formation mechanisms of different types of fringe-order noise. Then, an edge-preserved fringe-order correction strategy has been proposed to tackle the fringe order noise problem. Unlike the proprietary denoising algorithms, the correction approach is suitable for all types of CBFPP. Furthermore, the present algorithm is fast and can process in a short time. Moreover, the entire filtering procedure is noninvasive and can deal with dense point artifacts. The simulation and experimental results are presented to verify the feasibility and effectiveness of the proposed method.

基于代码的条纹投影轮廓仪（CBFPP）由于其简单的结构，坚固性和潜在的高速性而被广泛使用。此技术中最重要的步骤之一是正确获取条纹顺序。但是，条纹顺序往往很容易失效，尤其是在跳相区域附近。首先，本文分析了不同类型的条纹噪声的形成机理。然后，提出了一种边缘保留的边缘阶校正策略，以解决边缘阶噪声问题。与专有的降噪算法不同，该校正方法适用于所有类型的CBFPP。此外，本算法是快速的并且可以在短时间内处理。此外，整个过滤过程是非侵入性的，可以处理密集点伪像。