F constructing the of 213 buildings buildings as the reference constructing height information for the evaluation of heights. The reference reference location is shown in shown 1 beneath. 1 beneath. constructing heights. The creating constructing place is Figure in FigureFigure 1. GF-7 multi-spectral and multi-view image with the study area. Figure 1. GF-7 multi-spectral and multi-view image with the study location.three. Methodology 3. Methodology 3.1. Overview three.1. Overview The 3D information extraction approach in the developing in in this studyshown in Compound 48/80 In stock FigThe 3D facts extraction method on the developing this study is is shown in Figure Very first, we fused the GF-7 backward-view multi-spectral image together with the backwardure two. 2. Very first, we fused the GF-7 backward-view multi-spectral image with the backwardview panchromatic image and proposed MSAU-Net to extract the the urban developing footview panchromatic image and proposed MSAU-Net to extract urban creating footprint in the pan sharpening result. We modified the traditional decoder ncoder Icosabutate supplier network print from the pan sharpening result. We modified the conventional decoder ncoder netstructure, employed ResNet34 because the backbone function extraction network, andand integrated work structure, utilised ResNet34 because the backbone function extraction network, integrated an attention block in the skipskip connection component ofnetwork. The attention mechanism was an focus block inside the connection a part of the the network. The interest mechanism used utilized to enhance the building extraction ability in the neural network. Second, the was to improve the building extraction capability with the neural network. Second, the pointRemote Sens. 2021, 13, 4532 Remote Sens. 2021, 13, x FOR PEER Critique Remote Sens. 2021, 13, x FOR PEER REVIEW4 of 20 4 of 20 four ofcloud with the study area was constructed in the multi-view imagesimages ofand then point cloud from the study location was constructed from the multi-view of GF-7, GF-7, and point cloud the study region was constructed from on multi-view images of GF-7,utilized a study area as well as the DSM of from the the studywas constructed primarily based the the point cloud. Then, we we employed then the DSM of region was constructed depending on the point cloud. Then, then simulation the study region was DSM of algorithm (CSF) [34] to filter the point the point Then, we applied cloththe simulation algorithm (CSF)constructed based oncloud totocloud.the ground point a cloth [34] to filter the point cloud acquire the ground point get a cloth simulation algorithm (CSF) [34] filter the point cloud to get the constructed and applied itit to construct the DEM of to study area. Then, the nDSM wasground point toto to construct the DEM of the study location. Then, the nDSM was constructed and applied the and employed the height in the DEM objects. Ultimately, the building footprint extraction benefits towards the study area. Then, the nDSM was to represent it theconstructoff-terrain ofobjects. Finally, the developing footprintconstructedresults represent height of off-terrain extraction represent the height using the nDSM to generate building height. Within the accuracy assessment of off-terrain objects. Ultimately, the building footprint extraction results had been superimposed using the nDSM to produce building height. Inside the accuracy assesswere superimposed were superimposed together with the nDSM to create part of component study, study, the test dataset and thebuilding height. Within the accuracy assess- to ment our of our the test dataset as well as the reference developing height worth have been employed reference developing height.