Enhancing Three-Dimensional Reconstruction Through Intelligent Colormap Selection

Photogrammetry is the process of creating three-dimensional (3D) reconstructions from two-dimensional images. In this paper, a novel method is introduced that significantly enhances 3D reconstruction by improving image quality through a combination of RGB stretching, Contrast Limited Adaptive Histogram Equalization (CLAHE), and colormaps. This approach effectively increases the number of keypoints and matches between images, resulting in more accurate and detailed 3D reconstructions. Additionally, a heuristic methodology is proposed to identify the optimal colormaps for each dataset based on keypoint matches between image pairs. This approach saves valuable time by recommending the most effective colormaps, eliminating the need to test each individually. A new dataset comprising two real-world image collections from underwater cultural heritage sites is presented to validate the algorithm, characterized by challenging environments such as low visibility and varying lighting conditions. The limitations of existing techniques are addressed by this method, providing a robust solution for enhancing image quality in demanding underwater scenarios. Experimental results show that image enhancement can lead to a 7.91% and 11.4% improvement, and the enhancement fusion with colormaps improves reconstruction accuracy by 10.82% and 64.11%. These findings render the proposed methodology a promising tool for 3D reconstruction in challenging environments, like underwater.