Crop mapping is essential for agricultural management and food production monitoring, but challenges like limited crop labels and poor model generalization significantly hinder large-scale crop mapping. Here, we intro...
详细信息
Crop mapping is essential for agricultural management and food production monitoring, but challenges like limited crop labels and poor model generalization significantly hinder large-scale crop mapping. Here, we introduce a novel Contrastive Self-Training Network (CSTN), integrating a self-training strategy and contrastive domain adaptation (CDA) for cross-region crop mapping. CSTN uses pseudo-labels in the target region generated by the self-training strategy to assist supervised learning, and aligns features across regions using class-aware prototypes. Qualitative and quantitative evaluations demonstrate that CSTN significantly outperforms state-ofthe-art methods with a 12.29 % increase in average F1-score, particularly in maize identification. Moreover, CSTN also enables early-season crop classification for pre-harvest decision-making applications. The interpretability of the model is demonstrated through an in-depth analysis of feature map visualizations, attention map visualizations, and the effectiveness of the modules. This study provides a robust method for enhancing largescale crop mapping and facilitating more accurate and timely agricultural practices.
Removal of sulfur species from blast furnace gas is urgently needed due to the strict emission limits imposed on iron-steel industrial flue gas. Improving the sulfur capacity of H2S is a crucial challenge to reduce th...
详细信息
Removal of sulfur species from blast furnace gas is urgently needed due to the strict emission limits imposed on iron-steel industrial flue gas. Improving the sulfur capacity of H2S is a crucial challenge to reduce the operation cost. NiFe layered double hydroxide (LDH) adsorbents were synthesized using the hydrothermal method to strengthen the adsorption of H2S, achieving a high sulfur capacity of 133.6 mg/g at 50 degrees C. Characterization studies have revealed that the reaction pathway of H2S on the NiFe LDH surface involves adsorption, dissociation and oxidation. It has been clarified that the high sulfur capacity can be attributed to the abundant H2S dissociation sites and the excellent O2 activation sites. The dissociation sites of H2S encompass metal sites, -OH and CO32-. The interaction between O2 and the bridge site of asymmetric metal atoms significantly enhances the dissociation of O2. Strengthening the dissociation of H2S and O2 improves the sulfur capacity. The deactivation of adsorbents comes from the continuous consumption of oxygen species mainly composed of -OH and the deposition of sulfur species in the smaller mesopores ranging from 2 to 10 nm. This work provides useful insights into designing highly efficient iron-based adsorbents for the desulfurization of blast furnace gas.
Although good performance has been recently achieved in point cloud semantic segmentation based on deep learning, it has not been promoted due to differences in actual scenes and the time-consuming production of label...
详细信息
Although good performance has been recently achieved in point cloud semantic segmentation based on deep learning, it has not been promoted due to differences in actual scenes and the time-consuming production of labeled data sets. Unsupervised domain adaptation (UDA) aims to solve the problem of how to adapt the classifier from one scene (source domain) to another unlabeled scene (target domain), which can reduce the performance drop caused by the domain shift. Since spatial information is important for light detection and ranging (LiDAR) point and the causes of domain gap in image and point cloud tasks are different, projecting point cloud into image for processing is not suitable and how to apply the image-oriented domain adaptation (DA) methods to point cloud is not trivial. In this letter, we propose a 3D point-based UDA method for point cloud semantic segmentation. This model introduces point- and set-level domain adaptive modules to achieve feature alignment between the domains. We evaluate the proposed method with two experiments, including cross terrain adaptation and airborne to mobile adaptation. Compared with the results without using DA, the mean intersection over union (mIoU) increased by 10.45% and 24.69%, respectively, indicating the effectiveness of our method.
sAs the major maize-cultivated areas, the one-season cropland of China is increasingly threatened by rapid urbanization and soybean rejuvenation. Quantifying the area changes of maize cropland is crucial for both food...
sAs the major maize-cultivated areas, the one-season cropland of China is increasingly threatened by rapid urbanization and soybean rejuvenation. Quantifying the area changes of maize cropland is crucial for both food and energy security. Nonetheless, due to the lack of survey data related to planting types, long-term and fine-grained maize cropland maps in China dominated by small-scale farmlands are still unavailable. In this paper, we collect 75,657 samples based on field surveys and propose a deep learning-based method according to the phenology information of maize. With the generalization capability, the proposed method produces maize cropland maps with a resolution of 30 m from 2013 to 2021 in the one-season planting areas of China. The maize-cultivated areas derived from the maps are highly consistent with the data recorded by statistical yearbooks (R-2 = 0.85 on average), which indicates that the produced maps are reliable to facilitate the research on food and energy security.
China contributed nearly one-fifth of the world maize production over the past few years. Mapping the distributions of maize cropland in China is crucial to ensure global food security. Nonetheless, 10 m maize croplan...
China contributed nearly one-fifth of the world maize production over the past few years. Mapping the distributions of maize cropland in China is crucial to ensure global food security. Nonetheless, 10 m maize cropland maps in China are still unavailable, restricting the promotion of sustainable agriculture. In this paper, we collect numerous samples to produce annual 10-m maize cropland maps in China from 2017 to 2021 with a machine learning based classification framework. To overcome the temporal variations of plants, the proposed framework takes Sentinel-2 sequence images as input and utilizes deep neural networks and random forest as classifiers to map maize in a zone-specific way. The generated maps have an overall accuracy (OA) spanning from 0.87 to 0.95 and the maize-cultivated areas estimated by the maps are highly consistent with the records in statistical yearbooks (R2 varying from 0.83 to 0.95). To the best of our knowledge, this is the first annual 10-m maize maps across China, which largely facilitates the sustainable agriculture development in China dominated by smallholder farmlands.
Spaceborne full-waveform LiDAR has shown unique advantages in measuring global surface elevation. Laser footprints generally have decimeter-level vertical accuracy, meeting the requirement of ground elevation control ...
详细信息
Spaceborne full-waveform LiDAR has shown unique advantages in measuring global surface elevation. Laser footprints generally have decimeter-level vertical accuracy, meeting the requirement of ground elevation control points. In contrast, the footprint horizontal geolocation accuracy is in the meter to ten-meter levels. Although previous researches attempted to locate the footprint horizontal coordinate based on the digital surface model (DSM), the applicability and performance of the DSM-based positioning method in evaluating the footprint geolocation accuracy should be rigorously assessed before large-scale applications. Therefore, this study practices the DSM-based footprint positioning method over several study sites with various land covers and different laser campaigns. The footprint geolocation accuracy of the ICESat/GLAS (Ice, Cloud, and land Elevation Satellite/Geoscience Laser Altimeter System), the first Earth observation full-waveform LiDAR satellite, is evaluated by the DSM-based method. Results indicate that the DSM-based positioning method is only suitable for areas with significant height features, but not applicable in areas with high spatial correlation. The derived footprint geolocation accuracy (8.19-m horizontal shifting with 4.19-m standard deviation) is relatively reliable in urban site with relatively high spatial heterogeneity. This study helps make better use of the DSM-based footprint positioning method and design calibration experiments of full-waveform LiDAR satellites.
Accurate and timely maize yield monitoring from satellite imagery is in great demand in developing countries. The spatial heterogeneity deprived of the large territory of China makes it a challenge. In this article, w...
详细信息
Accurate and timely maize yield monitoring from satellite imagery is in great demand in developing countries. The spatial heterogeneity deprived of the large territory of China makes it a challenge. In this article, we developed a novel deep learning model for maize yield prediction at the county level based on multiple satellite data. The two-stage feature learning structure integrated data from disparate sources, enhancing feature representation. The discriminative features were optimized from two levels: the dictionary matrix learned the spatial diversity of the provinces, and the improved optimizing formulation fit the distribution of the unbalanced records. The cross-validation results showed that our approach could explain 82 % of the variation in maize yield, achieving state-of-the-art. The model was robust when predicting the future, with the average root-mean-square error of 1006 kg/ha and the mean-absolute-percentage error of 17.1 %. The ability of early maize yield prediction clarifies the tremendous application value, showing the data from the first two months can already explain 75.6 % of yield variation. It was the first effort to improve county-level maize yield prediction in China, providing a potential framework for advancing the use of multi-source datasets for maize yield estimating.
暂无评论