The automated detection of buildings in remote sensing images enables understanding the distribution information of buildings, which is indispensable for many geographic and social applications, such as urban planning, change monitoring and population estimation. The performance of deep learning in images often depends on a large number of manually labeled samples, the production of which is time-consuming and expensive. Thus, this study focuses on reducing the number of labeled samples used and proposing a semi-supervised deep learning approach based on an edge detection network (SDLED), which is the first to introduce semi-supervised learning to the edge detection neural network for extracting building roof boundaries from high-resolution remote sensing images. This approach uses a small number of labeled samples and abundant unlabeled images for joint training. An expert-level semantic edge segmentation model is trained based on labeled samples, which guides unlabeled images to generate pseudo-labels automatically. The inaccurate label sets and manually labeled samples are used to update the semantic edge model together. Particularly, we modified the semantic segmentation network D-LinkNet to obtain high-quality pseudo-labels. Specifically, the main network architecture of D-LinkNet is retained while the multi-scale fusion is added in its second half to improve its performance on edge detection. The SDLED was tested on high-spatial-resolution remote sensing images taken from Google Earth. Results show that the SDLED performs better than the fully supervised method. Moreover, when the trained models were used to predict buildings in the neighboring counties, our approach was superior to the supervised way, with line IoU improvement of at least 6.47% and F1 score improvement of at least 7.49%.

Thin clouds in remote sensing images increase the radiometric distortion of land surfaces. The identification of pixels contaminated by thin clouds, known as the thin-cloud mask, is an important preprocessing procedure to guarantee the proper utilization of data. However, failure to effectively separate thin clouds and high-reflective land-cover features causes thin-cloud masks to remain a challenge. To overcome this problem, we developed a thin-cloud masking method for remote sensing images based on sparse dark pixel region detection. As a result of the effect of scattering, the path radiance is added to the radiance recorded by the sensor in the thin-cloud area, which causes the number of dark pixels in the thin-cloud area to be much less than that in the clear area. In this study, the area of a Thiessen polygon (a nonparametric measure) is used to evaluate the density of local dark pixels, and the region with the sparse dark pixel is selected as the thin-cloud candidate. Then, thin-cloud and clear areas are used as samples to train the background suppression haze thickness index (BSHTI) transform parameters, and convert the original multiband images into single-band images. Finally, an accurate thin-cloud mask is obtained for every buffered thin-cloud candidate, via the segmentation of the BSHTI band. Additionally, the multispectral images obtained by the Wide Field View (WFV), on board the Chinese GaoFen1, and the Operational Land Imager (OLI), on board the Landsat 8, are employed to evaluate the performance of the method. The results reveal that the proposed approach can obtain a thin-cloud mask with a high true-value ratio and detection ratio. Thin-cloud masks can satisfy various application demands.

Building change detection (BuCD) can offer fundamental data for applications such as urban planning and identifying illegally-built new buildings. With the development of deep neural network-based approaches, BuCD using high-spatial-resolution remote sensing images (RSIs) has significantly advanced. These deep neural network-based methods, nevertheless, typically demand a considerable number of computational resources. Additionally, the accuracy of these algorithms can be improved. Hence, LightCDNet, a lightweight Siamese neural network for BuCD, is introduced in this paper. Specifically, LightCDNet comprises three components: a Siamese encoder, a multi-temporal feature fusion module (MultiTFFM), and a decoder. In the Siamese encoder, MobileNetV2 is chosen as the feature extractor to decrease computational costs. Afterward, the multi-temporal features from dual branches are independently concatenated based on the layer level. Subsequently, multiscale features computed from higher levels are up-sampled and fused with the lower-level ones. In the decoder, deconvolutional layers are adopted to gradually recover the changed buildings. The proposed network LightCDNet was assessed using two public datasets: namely, the LEVIR BuCD dataset (LEVIRCD) and the WHU BuCD dataset (WHUCD). The F1 scores on the LEVIRCD and WHUCD datasets of LightCDNet were 89.6% and 91.5%, respectively. The results of the comparative experiments demonstrate that LightCDNet outperforms several state-of-the-art methods in accuracy and efficiency.

Building extraction is a basic task in the field of remote sensing, and it has also been a popular research topic in the past decade. However, the shape of the semantic polygon generated by semantic segmentation is irregular and does not match the actual building boundary. The boundary of buildings generated by semantic edge detection has difficulty ensuring continuity and integrity. Due to the aforementioned problems, we cannot directly apply the results in many drawing tasks and engineering applications. In this paper, we propose a novel convolutional neural network (CNN) model based on multitask learning, Dense D-LinkNet (DDLNet), which adopts full-scale skip connections and edge guidance module to ensure the effective combination of low-level information and high-level information. DDLNet has good adaptability to both semantic segmentation tasks and edge detection tasks. Moreover, we propose a universal postprocessing method that integrates semantic edges and semantic polygons. It can solve the aforementioned problems and more accurately locate buildings, especially building boundaries. The experimental results show that DDLNet achieves great improvements compared with other edge detection and semantic segmentation networks. Our postprocessing method is effective and universal.

Edge detection of ground objects is a typical task in the field of remote sensing and has advantages in accomplishing many complex ground object extraction tasks. Although recent mainstream edge detection methods based on deep learning have significant effects, these methods have a very high dependence on the quantity and quality of samples. Moreover, using datasets from other domains in detection tasks often leads to degraded network performance due to variations in the ground objects in different regions. If this problem can be solved to allow datasets from other domains to be reused, the number of labeled samples required in the new task domain can be reduced, thereby shortening the task cycle and reducing task costs. In this paper, we propose a weakly supervised domain adaptation method to address the high dependence of edge extraction networks on samples. The domain adaptation is performed on the edge level and the semantic level, which prevents deviations in the semantic features that are caused by the overgeneralization of edge features. Additionally, the effectiveness of our proposed domain adaptation module is verified. Finally, we demonstrate the superior edge extraction performance of our method in the SEGOS edge extraction network in contrast to other edge extraction methods.

Deep convolutional neural network (DCNN)-based methods have shown great improvements in building extraction from high spatial resolution remote sensing images. In this paper, we propose a postprocessing method based on DCNNs for building extraction. Specifically, building regions and boundaries are learned simultaneously or separately by DCNNs. The predicted building regions and boundaries are then combined by the postprocessing method to produce the final building regions. In addition, we introduce a manually labeled dataset based on high spatial resolution images for building detection, the XIHU building dataset. This dataset is then used in the experiments to evaluate our methods. Besides the WHU building dataset, East Asia (WHUEA) is also included. Results demonstrate that our method that combines the results of DeepLab and BDCN shows the best performance on the XIHU building dataset, which achieves 0.78% and 23.30% F1 scores, and 1.13% and 28.45% intersection-over-union (IoU) improvements compared with DeepLab and BDCN, respectively. Additionally, our method that combines the results of Mask R-CNN and DexiNed performs best on the WHUEA dataset. Moreover, our methods outperform the state-of-the-art multitask learning network, PMNet, on both XIHU and WHUEA datasets, which indicates that the overall performance can be improved although building regions and boundaries are learned in the training stage.