Authors: Jian Liang*, School of Resource and Environment Sciences, Wuhan University, 129 Luoyu Road, Wuhan 430079, China, Lin Li, School of Resource and Environment Sciences, Wuhan University, 129 Luoyu Road, Wuhan 430079, China, Haihong Zhu, School of Resource and Environment Sciences, Wuhan University, 129 Luoyu Road, Wuhan 430079, China
Topics: Remote Sensing
Keywords: “building extraction”, “remote sensing”, “deep learning”, “CNN”
Session Type: Paper
Start / End Time: 2:35 PM / 4:15 PM
Room: Virginia B, Marriott, Lobby Level
Presentation File: No File Uploaded
Automatic building extraction from remote sensing imagery is important in many applications. The success of convolutional neural networks (CNNs) has also led to advances in using CNNs to extract man-made objects from high-resolution imagery. However, the large appearance and size variations of buildings make it difficult to extract both crowded small buildings and large buildings. High-resolution imagery must be segmented into patches for CNN models due to GPU memory limitations, and buildings are typically only partially contained in a single patch with little context information. To overcome the problems involved when using different levels of image features with common CNN models, this paper proposes a novel CNN architecture called a multiple-feature reuse network in which each layer is connected to all the subsequent layers of the same size, enabling the direct use of the hierarchical features in each layer. In addition, the model includes a smart decoder that enables precise localization with less GPU load. We tested our model on the large real-world remote sensing datasets and obtained the results with higher accuracy and score than that by CNN models. The experiments show that the proposed model with connecting convolutional layers improves the performance of common CNN models for extracting buildings of different sizes and can output satisfactory results at a consumer-level GPU.