Search

Change detection is the process of identifying pixelwise differences in bitemporal co-registered images. It is of great significance to Earth observations. Recently, with the emergence of deep learning (DL), the power and feasibility of deep convolutional neural network (CNN)-based methods have been shown in the field of change detection. However, there is still a lack of effective supervision for change feature learning. In this work, a feature constraint change detection network (FCCDN) is proposed. We constrain features both in bitemporal feature extraction and feature fusion. More specifically, we propose a dual encoder-decoder network backbone for the change detection task. At the center of the backbone, we design a nonlocal feature pyramid network to extract and fuse multiscale features. To fuse bitemporal features in a robust way, we build a dense connection-based feature fusion module. Moreover, a self-supervised learning-based strategy is proposed to constrain feature learning. Based on FCCDN, we achieve state-of-the-art performance on two building change detection datasets (LEVIR-CD and WHU). On the LEVIR-CD dataset, we achieve an IoU of 0.8569 and an F1 score of 0.9229. On the WHU dataset, we achieve an IoU of 0.8820 and an F1 score of 0.9373. Moreover, for the first time, the acquisition of accurate bitemporal semantic segmentation results is achieved without using semantic segmentation labels. This is vital for the application of change detection because it saves the cost of labeling., 46 pages, 16 figures. Submitted to ISPRS Journal of Photogrammetry and Remote Sensing. Under review

Semantic segmentation is an essential part of deep learning. In recent years, with the development of remote sensing big data, semantic segmentation has been increasingly used in remote sensing. Deep convolutional neural networks (DCNNs) face the challenge of feature fusion: very-high-resolution remote sensing image multisource data fusion can increase the network's learnable information, which is conducive to correctly classifying target objects by DCNNs; simultaneously, the fusion of high-level abstract features and low-level spatial features can improve the classification accuracy at the border between target objects. In this paper, we propose a multipath encoder structure to extract features of multipath inputs, a multipath attention-fused block module to fuse multipath features, and a refinement attention-fused block module to fuse high-level abstract features and low-level spatial features. Furthermore, we propose a novel convolutional neural network architecture, named attention-fused network (AFNet). Based on our AFNet, we achieve state-of-the-art performance with an overall accuracy of 91.7% and a mean F1 score of 90.96% on the ISPRS Vaihingen 2D dataset and an overall accuracy of 92.1% and a mean F1 score of 93.44% on the ISPRS Potsdam 2D dataset., Comment: 35 pages. Published by ISPRS Journal of Photogrammetry and Remote Sensing

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.

The land-use mapping of Beijing-Tianjin-Hebei region (Jingjinji) covers an area of over 200,000 km2. Faced with a large number of objects in tiny size, which is quite difficult for land-use mapping, the land-use mapping of Jingjinji requires a certain design of convolutional neural network (CNN) [1] (Lecun et al. Neural Comput 1(4):541–551, [2]. In this paper, we propose a feature pyramid fusion network with attention mechanism. The module is designed based on the pyramid scene parsing network (PSPNet) (Zhao et al. [3]), and two improvements are made for the large-scale high-resolution land-use mapping task: (1) Since PSPNet concatenate features in a non-selective way at the center of the module, we propose the attention feature pyramid fusion (AFPF) block, which can selectively fuse features with different scales; (2) In order to make the border of results more precise and the tiny objects more accurate, we use an encode–decode structure to merge low-level features and high-level features at the upsample stage. During each fusion operation, the attention mechanism is employed. The final experiment proves that the network designed in this paper performs better in the accuracy of small objects and precise border with respect to the original PSPNet in the land-use mapping task of Jingjinji.

Ensuring food security is a long-term and arduous task. Timely and accurate grasp of grain production capacity information can provide favourable data support for the nation to formulate macroeconomic plans and food policies. With the development of remote sensing technology, it has been widely used in crop yield estimation models. In this paper, the yield of spring maize in Da’an of Jilin province was estimated based on vegetation indexes calculated from Landsat-8 images. The results have shown that the fitting degree and estimation accuracy of yield estimation models at tasselling stage are significantly better than those at milk stage. Among these vegetation indexes, the model based on GNDVI has better fitting degree and estimation accuracy. This paper can provide reference for the post construction evaluation of high standard farmland in China.

Object detection is facing various challenges as an important aspect in the field of remote sensing&mdash, especially in large scenes due to the increase of satellite image resolution and the complexity of land covers. Because of the diversity of the appearance of track and fields, the complexity of the background and the variety between satellite images, even superior deep learning methods have difficulty extracting accurate characteristics of track and field from large complex scenes, such as the whole of China. Taking track and field as a study case, we propose a stable and accurate method for target detection. Firstly, we add the &ldquo, deconvolution&rdquo, and &ldquo, concat&rdquo, module to the structure of the original Single Shot MultiBox Detector (SSD), where Visual Geometry Group 16 (VGG16) is served as a basic network, followed by multiple convolution layers. The two modules are used to sample the high-level feature map and connect it with the low-level feature map to form a new network structure multi-scale-fused SSD (abbreviated as MSF_SSD). MSF-SSD can enrich the semantic information of the low-level feature, which is especially effective for small targets in large scenes. In addition, a large number of track and fields are collected as samples for the whole China and a series of parameters are designed to optimize the MSF_SSD network through the deep analysis of sample characteristics. Finally, by using MSF_SSD network, we achieve the rapid and automatic detection of meter-level track and fields in the country for the first time. The proposed MSF_SSD model achieves 97.9% mean average precision (mAP) on validation set which is superior to the 88.4% mAP of the original SSD. Apart from this, the model can achieve an accuracy of 94.3% while keeping the recall rate in a high level (98.8%) in the nationally distributed test set, outperforming the original SSD method.

The timely and accurate mapping and monitoring of mine tailings dams is crucial to the improvement of management practices by decision makers and to the prevention of disasters caused by failures of these dams. Due to the complex topography, varying geomorphological characteristics, and the diversity of ore types and mining activities, as well as the range of scales and production processes involved, as they appear in remote sensing imagery, tailings dams vary in terms of their scale, color, shape, and surrounding background. The application of high-resolution satellite imagery for automatic detection of tailings dams at large spatial scales has been barely reported. In this study, a target detection method based on deep learning was developed for identifying the locations of tailings ponds and obtaining their geographical distribution from high-resolution satellite imagery automatically. Training samples were produced based on the characteristics of tailings ponds in satellite images. According to the sample characteristics, the Single Shot Multibox Detector (SSD) model was fine-tuned during model training. The results showed that a detection accuracy of 90.2% and a recall rate of 88.7% could be obtained. Based on the optimized SSD model, 2221 tailing ponds were extracted from Gaofen-1 high resolution imagery in the Jing–Jin–Ji region in northern China. In this region, the majority of tailings ponds are located at high altitudes in remote mountainous areas. At the city level, the tailings ponds were found to be located mainly in Chengde, Tangshan, and Zhangjiakou. The results prove that the deep learning method is very effective at detecting complex land-cover features from remote sensing images.

The land-use map is an important data that can reflect the use and transformation of human land, and can provide valuable reference for land-use planning. For the traditional image classification method, producing a high spatial resolution (HSR), land-use map in large-scale is a big project that requires a lot of human labor, time, and financial expenditure. The rise of the deep learning technique provides a new solution to the problems above. This paper proposes a fast and precise method that can achieve large-scale land-use classification based on deep convolutional neural network (DCNN). In this paper, we optimize the data tiling method and the structure of DCNN for the multi-channel data and the splicing edge effect, which are unique to remote sensing deep learning, and improve the accuracy of land-use classification. We apply our improved methods in the Guangdong Province of China using GF-1 images, and achieve the land-use classification accuracy of 81.52%. It takes only 13 hours to complete the work, which will take several months for human labor., Comment: Accepted at IEEE International Geoscience and Remote Sensing Symposium (IGARSS) 2019