Convolutional Neural Network for Detection of Building Contours Using Multisource Spatial Data

Building reconstruction from aerial photographs and other multi-source urban spatial data is a task endeavored using a plethora of automated and semi-automated methods ranging from point processes, classic image processing and laser scanning. Here, we describe a convolutional neural network (CNN) method for the detection of building borders. In particular, the network is based on the state of the art super-resolution model SRCNN and accepts aerial photographs depicting densely populated urban area data as well as their corresponding digital elevation maps (DEM). Training is performed using three variations of this urban data set and aims at detecting building contours through a novel super-resolved heteroassociative mapping. Another novelty of our approach is the design of a modified custom loss layer, named Top-N, whereby the mean square error (MSE) between the reconstructed output image and the provided ground truth (GT) image of building contours is computed on the 2N image pixels with highest values, where N is the number of contour pixels in GT. Assuming that most of the N contour pixels of the GT image are also in the top 2N pixels of the reconstruction, this modification balances the two pixel categories and improves the generalization behavior of the CNN model. It is shown, in our experiments, that the Top-N cost function offers performance gains in comparison to standard MSE. Further improvement in generalization ability of the network is achieved by using dropout.