Change detection (CD) based on high-resolution remote sensing images can be used to monitor land cover changes, which is an important and challenging topic in the remote sensing field. In recent years, with the development of deep learning, CD methods based on deep learning have achieved good results in the field of CD. However, most current CD methods use single-stream or dual-stream networks to extract change features, which is insufficient to extract and learn bitemporal change information thoroughly. This paper proposes a triple-stream network (TSNet) with cross-stage feature fusion for CD in high-resolution bitemporal remote sensing images. First, to obtain highly representative deep features in the original image, we perform feature extraction on bitemporal remote sensing images and their concatenated image with a dual-stream encoder and a single-stream encoder, respectively. Then, the bitemporal multiscale features extracted by the dual-stream encoder are input into a multistage bidirectional convolutional gated recurrent unit (MSBC_GRU) feature fusion module, allowing the network to learn the change information in a cross-stage manner. In addition, we use a dual-channel attention module to fuse the features extracted by dual-stream and single-stream encoders, improving the network’s ability to discriminate changed features. The effectiveness of TSNet is demonstrated with three publicly available CD datasets. The extensive experimental results demonstrate that the proposed method achieves state-of-the-art CD performance on the above three datasets.

Deep learning (DL)-based architectures have shown a strong capacity to identify changes. However, existing change detection (CD) networks still suffer from limited applicability when it comes to multi-scale targets and spatially misaligned objects. For the sake of tackling the above problems, a region-based feature fusion network (RFNet) for CD of very high spatial resolution (VHR) remote sensing images is proposed. RFNet uses a fully convolutional Siamese network backbone where a multi-stage feature interaction module (MFIM) is embedded in the dual encoder and a series of region-based feature fusion modules (RFFMs) is used to generate change information. The MFIM fuses features in different stages to enhance the interaction of multi-scale information and help the network better distinguish complex ground objects. The RFFM is built based on region similarity (RSIM), which measures the similarity of bitemporal features with neighborhoods. The RFFM can reduce the impact of spatially offset bitemporal targets and accurately identify changes in bitemporal images. We also design a deep supervise strategy by directly introducing RSIM into loss calculation and shortening the error propagation distance. We validate RFNet with two popular CD datasets: the SECOND dataset and the WHU dataset. The qualitative and quantitative comparison results demonstrate the high capacity and strong robustness of RFNet. We also conduct robustness experiments and the results demonstrate that RFNet can deal with spatially shifted bitemporal images.

Deep learning-based semantic segmentation technology is widely applied in remote sensing and has achieved excellent performance in remote sensing image target extraction. Greenhouses play an important role in the development of agriculture in China. However, the rapid expansion of greenhouses has had a series of impacts on the environment. Therefore, the extraction of large-scale greenhouses is crucial for the sustainable development of agriculture and environmental governance. It is difficult for existing methods to acquire precise boundaries. Therefore, we propose a spatial convolutional long short-term memory structure, which can fully consider the spatial continuity of ground objects. We use multitask learning to improve the network’s ability to extract image boundaries and promote convergence through auxiliary loss. We propose a superpixel optimization module to optimize the main-branch results of network semantic segmentation using more precise boundaries obtained by advanced superpixel segmentation techniques. Compared with other mainstream methods, our proposed structure can better consider spatial information and obtain more accurate results. We chose Shandong Province, China, as the study area and used Gaofen-1 satellite remote sensing images to create a new greenhouse dataset. Our method achieved an F1 score of 77%, a significant improvement over mainstream semantic segmentation networks, and it could extract greenhouse results with more precise boundaries. We also completed large-scale greenhouse mapping for Shandong Province, and the results show that our proposed modules have great potential in greenhouse extraction.

The ecological redline defines areas where industrialization and urbanization development should be prohibited. Its purpose is to establish the most stringent environmental protection system to meet the urgent needs of ecological function guarantee and environmental safety. Nowadays, deep learning methods have been widely used in change detection tasks based on remote sensing images, which can just be applied to the monitoring of the ecological redline. Considering the convolution-based neural networks’ lack of utilization of global information, we choose a transformer to devise a Siamese network for change detection. We also use a transformer to design a pyramid pooling module to help the network maintain more features. Moreover, we construct a self-supervised network based on a contrastive method to obtain a pre-trained model, especially for remote sensing images, aiming to achieve better results. As for study areas and data sources, we chose Hebei Province, where the environmental problem is quite nervous, and used its GF-1 satellite images to do our research. Through ablation experiments and contrast experiments, our method is proven to have significant advantages in terms of accuracy and efficiency. We also predict large-scale areas and calculate the intersection recall rate, which confirms that our method has practical values.

With the development of deep learning, semantic segmentation technology has gradually become the mainstream technical method in large-scale multi-temporal landcover classification. Large-scale and multi-temporal are the two significant characteristics of Landsat imagery. However, the mainstream single-temporal semantic segmentation network lacks the constraints and assistance of pre-temporal information, resulting in unstable results, poor generalization ability, and inconsistency with the actual situation in the multi-temporal classification results. In this paper, we propose a multi-temporal network that introduces pre-temporal information as prior constrained auxiliary knowledge. We propose an element-wise weighting block module to improve the fine-grainedness of feature optimization. We propose a chained deduced classification strategy to improve multi-temporal classification’s stability and generalization ability. We label the large-scale multi-temporal Landsat landcover classification dataset with an overall classification accuracy of over 90%. Through extensive experiments, compared with the mainstream semantic segmentation methods, our proposed multi-temporal network achieves state-of-the-art performance with good robustness and generalization ability.

In recent years, with the development of deep learning in remotely sensed big data, semantic segmentation has been widely used in large-scale landcover classification. Landsat imagery has the advantages of wide coverage, easy acquisition, and good quality. However, there are two significant challenges for the semantic segmentation of mid-resolution remote sensing images: the insufficient feature extraction capability of deep convolutional neural network (DCNN); low edge contour accuracy. In this paper, we propose a block shuffle module to enhance the feature extraction capability of DCNN, a differentiable superpixel branch to optimize the feature of small objects and the accuracy of edge contours, and a self-boosting method to fuse semantic information and edge contour information to further optimize the fine-grained edge contour. We label three sets of Landsat landcover classification datasets, and achieved an overall accuracy of 86.3%, 83.2%, and 73.4% on the three datasets, respectively. Compared with other mainstream semantic segmentation networks, our proposed block shuffle network achieves state-of-the-art performance, and has good generalization ability.