在3D重建领域中，灵活且高效的网格生成对几何优化和逆向渲染等任务至关重要。然而，传统的等面值提取方法在处理复杂几何结构时，常常存在精度不足、拓扑不稳定和自相交等问题，限制了其在高精度任务中的应用。为了解决这些问题，提出了一种结合曲率与拓扑约束的感知优化方法（curvature and topology-constrained mesh optimization, CTMO）。该方法通过引入曲率约束，在高曲率区域动态调整网格形变，确保几何细节的精确保留。同时，使用拓扑约束来防止自交和不规则拓扑结构的形成。此外，还引入了自适应网格分辨率调整机制，以提高计算效率和优化精度。实验结果表明，该方法显著提高了CTMO在处理复杂几何体时的稳定性和精度，使其适用于各种3D重建任务。研究结果为高精度3D重建提供了一种新方法，并扩展了灵活网格生成技术的应用范围。

In the field of 3D reconstruction, flexible and efficient mesh generation is essential for tasks such as geometry optimization and inverse rendering. Traditional isosurface extraction methods often suffer from issues such as insufficient accuracy, topological instability, and self-intersections when processing complex geometric structures, limiting their application in high-precision tasks. To address these challenges, a curvature-aware optimization method combined with topological constraints CTMO(curvature and topology-constrained mesh optimization,) was proposed. The CTMO method dynamically adjusts mesh deformation in high-curvature regions by introducing curvature constraints, ensuring the accurate preservation of geometric details. Topological constraints were also applied to prevent self-intersections and the formation of irregular topological structures. Additionally, an adaptive mesh resolution adjustment mechanism was incorporated to enhance computational efficiency and optimization precision. Experimental results demonstrate that this method significantly improves the stability and accuracy of CTMO when processing complex geometries, making it suitable for a wide range of 3D reconstruction tasks. This work offers a novel approach to high-precision 3D reconstruction and expands the application of flexible mesh generation techniques.