三维重建中纹理图像的光照一致性调整

A Review of Illuminance Consistency Adjustment of Texture Images for 3D Reconstruction

三维激光测量技术(LMS)为快速、精确的建立实体三维数字模型提供了新的途径。纹理图像的光照调整作为其中的一个关键技术环节,其目的是消除多幅纹理图像由于拍摄光照不一致造成的亮度不一致,从而增强模型的视觉真实感。在大型户外实体的三维重建过程中该问题尤为突出,但目前仍没有成熟的解决方案。文中在探讨问题成因、对基本问题进行描述的基础上,由浅入深的总结归纳了当前三类典型算法,探讨了各算法的原理、结合实验结果分析利弊、并对今后的研究进行了展望,以期为LMS技术的实际应用提供技术支撑。

3D laser measurement technology（LMS） provides a novel approach for rapidly and accurately recon- structing 3D digital model of an object in real world, llluminance consistency adjustment of texture images, as one of the key tasks, is aimed at eliminating the illuminance difference among multiple texture images, so as to enhance model＇s visual reality. The problem is significant especially for digitalization of outdoor objects in large - scale, but not solved perfectly yet. On the basis of a description of this problem, three types of algorithms are summarized, with their theory analyzed and their merits and drawbacks compared according to experimental results, and further research is also prospected, which is expected to be technically supportive for application based on LMS.