Aerospace mold profile image feature optimization method based on trajectory planning

This paper addresses multi-scale feature imbalance and local detail blurring in aerospace Invar steel mold cross-sectional images. It proposes a method that couples a multi-scale pyramid with an edge-preserving mechanism to enhance image features. We analyze the imbalance of multi-scale characteristics and the blurring of local details in aerospace mold profiles. Global and local feature extraction is performed using a multi-scale Gaussian pyramid. We introduce brightness compensation and dynamic range limiting mechanisms and implement recursive fusion to optimize image quality. Edge features are extracted using Laplace filtering, and edge saliency is enhanced by nonlinear gamma correction. This optimization improves edge and detail performance, with final edge protection provided by an edge-holding mechanism. The filtered peak signal-to-noise ratio reaches 40, the structural similarity index reaches 0.93, the mean square error is 0.0062, and the quality index approaches 1. The method significantly improves the peak signal-to-noise ratio and structural similarity index, enhances image edge definition and detailed features, and provides high quality for subsequent geometric reconstruction and depth computation.