【摘要】 变曲率复杂型面成形过程偏差评估对于确保其最终成形精度至关重要。现有成形偏差评估方法无法有效实现无明显几何特征的变曲率连续复杂型面目标点云与测量点云的精确匹配，且存在评估效率低、缺乏有效评估标准的问题。为此提出一种基于等距映射的变曲率复杂型面成形偏差评估方法，基于边界识别和等距映射实现测量点云与目标点云的精确配准，通过统计精配准后测量点云与标准点云各点分布，评估曲面成形偏差。基于法向量估计提取点云边界，对边界点应用ICP算法间接实现整体点云的粗配准，通过等距映射(ISOMAP)方法将测量点云和标准点云降维展平获取几何对应点对，通过求解配准目标函数实现点云的精配准，精配准后几何对应点对间距为曲面成形偏差评估。基于理论与技术研究成果，对激光喷丸成形后得到变曲率壁板进行成形偏差评估实验验证，结果表明，针对具有变曲率的复杂曲面，基于等距映射的变曲率复杂型面成形偏差评估方法显著提高了曲面偏差评估的精度和稳定性，且由于该方法无需迭代求解，算法的计算效率得到了提升，能更好地满足复杂型面构件在成形偏差评估中的需求。更多还原

【Abstract】 The assessment of forming deviations for complex surfaces with variable curvature is critical to ensuring their final forming accuracy. Existing methods fail to achieve precise matching between the target and measured point clouds for continuous complex surfaces with variable curvature and no prominent geometric features. These methods also suffer from low efficiency and a lack of effective evaluation standards. To address these issues, this paper proposes a forming deviation assessment method for complex surfaces with variable curvature based on ISOMAP. Accurate registration between the measured and target point clouds is achieved through boundary recognition and ISOMAP. The surface forming deviations are evaluated by statistically analyzing the point distributions of the measured and standard point clouds after fine registration. Boundary points are extracted using normal vector estimation, and the ICP algorithm is applied to these points for coarse registration of the overall point cloud. The ISOMAP method is then employed to reduce the dimensionality of the measured and standard point clouds, flattening them to obtain geometric corresponding point pairs. Fine registration is achieved by solving the registration objective function, and the distances between the geometric corresponding point pairs are used to assess the forming deviations. Theoretical and technical research findings are validated through experiments on forming deviation assessment of variable curvature wall panels obtained via laser peening forming. The results demonstrate that the proposed method significantly enhances the accuracy and stability of deviation assessments for complex surfaces with variable curvature. Furthermore, as the method eliminates the need for iterative solving, the computational efficiency is improved, meeting the demands for forming deviation assessment of complex surface components.更多还原