Spatial-Spectral Information Fusion Method and Application Based on Multispectral Point Cloud

Limited by imaging technology, multi-spectral imaging systems cannot obtain spatial information, which limits the application of quantitative remote sensing, such as biomass estimation, vegetation growth analysis, and natural resource measurement. To address the problem, a spatial-spectral information fusion method is proposed in this paper. Firstly, the Structure from Motion (SfM) algorithm is used to estimate the image and camera parameters, and then the image depth map point cloud is estimated by multi-view dense matching. Finally, the multi-spectral depth map point cloud is obtained by forward projection and fused into a multi-spectral dense point cloud. The experimental results show that the proposed method can reconstruct the 3D information of the observation scene from multi-spectral images, with high spatial geometric accuracy and complete spectral bands.