We develop a fully automatic method to reconstruct the LoD1 building from multi-view aerial images. The pipeline consists of three parts. (a) A novel deep learning-based multi-view matching method, composed of a convolutional neural network, gated recurrent convolutions, and a multi-scale pyramid matching structure, is used to reconstruct the digital surface model (DSM) and digital orthophoto map (DOM) efficiently without generating epipolarly rectified images. (b)A three-stage 2D building extraction method is used to deliver reliable and accurate building contours. Deep-learning based segmentation, assisted with DSM, is used to segment buildings from backgrounds; and the generated building maps are fused with a terrain classification algorithm to reach better segmentation results. A polygon regularization algorithm and a level set algorithm are thereafter employed to transfer the binary segmentation maps to structured vector-form building polygons. (c) A novel method is introduced to infer the height of building roofs and bases using adaptive local terrain filtering and neighborhood buffer analysis. The main core code is provided in this page.
Citing us if you find it is helpful in your study : Yu D , Ji S , Liu J , et al. Automatic 3D building reconstruction from multi-view aerial images with deep learning[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2021, 171(2021):155-170.
Download the network code. Click HERE to Download
Download the network code. Click HERE to Download