最新刊期
卷 26 ， 期 7 ， 2022
-
摘要：Lake ice is an important variable in cryospheric hydrology that is quite sensitive to climate change, serving as a regulator to regional climate as well as lake ecosystems and a natural resource to sustain human activities on the ice. Lake ice cover and Lake Ice Thickness (LIT) are two key variables in lake ice studies. However, most lake ice studies focus on lake ice cover as opposed to LIT, due mostly to the lack of in situ LIT observations and effective satellite platforms. Consequently, LIT states in ungauged basins remain largely unknown, and critical knowledge gaps need to be filled with breakthroughs in remote sensing observations and retrieval methods. This study reviews the progress in the remote sensing of LIT for the past two decades, focusing on methods based on active and passive microwave information as well as methods based on thermal infrared information. Currently, the best retrieval accuracy of LIT is 0.1—0.2 m for satellite platforms, unable to meet the accuracy requirement (1—2 cm) from the Global Climate Observing System (GCOS). Passive microwave sensors such as AMSR-E and AMSR2 with high temporal-resolution and coarse spatial resolution (~30 km) are efficient for monitoring LIT in large lakes. SAR images can be used to discriminate bedfast lake ice and floating ice, applicable to shallow lakes and lakes with bathymetry information. Waveforms of satellite altimeters containing signals backscattered from the lake ice surface and the lake ice bottom can be used to estimate the traveling time of radar pulse within the ice, which is suitable for ungauged basins because the method is physically based and does not need calibration. Backscattering coefficients from satellite altimeters can also reflect LIT evolution but needs further investigation to improve its robustness. Thermal infrared information such as MODIS land surface temperatures can be used to drive lake ice models to estimate the LIT, but its accuracy is subject to the complicated process of lake surface snow. We summarize the challenges for current LIT studies and indicate future directions for LIT studies. There are four critical issues: (1) The physical process of lake ice and lake surface snow needs to be better resolved to improve the performance of remote sensing and modeling of LIT; (2) In situ measurements of LIT need to be aggregated and standardized in a larger scale to provide sufficient calibration and validation data; (3) Multisource remote sensing information of LIT can be better combined with lake ice modeling through data fusion and data assimilation to derive global LIT datasets with a longer period, higher spatiotemporal resolution and better accuracy; (4) With the development of new generations of satellite platforms, such as the latest altimetry missions (e.g., ICESat-2, Jason-CS, and SWOT) and SAR platforms, advanced methods can be developed to better achieve the LIT monitoring target set by GCOS.关键词：lake ice thickness;brightness temperature;Backscattering coefficient;satellite altimetry;thermal infrared remote sensing- 1649
- |
- 533
- |
- 0
发布时间：2022-08-08
Review
-
摘要：The simulation and prediction model of land subsidence based on traditional numerical methods requires a large amount of hydrogeological and measured data, and predicting the deformation in areas with complex geological conditions is difficult. In this study, on the basis of land subsidence information obtained by permanent scatterers–interferometry synthetic aperture radar (PS-InSAR) technology in the east of the Beijing plain and in consideration of the influence of groundwater level in different layers on subsidence, the long-term and short-term memory network (AM-LSTM) based on an attention mechanism is used to simulate the land subsidence at typical locations in different subsidence areas. Results show the following points. (1) The spatial difference of land subsidence in the study area is obvious. From October 2010 to August 2016, the maximum subsidence rate is about 153 mm/a, and the cumulative subsidence is 1063 mm. The area is located near Sanjianfang Township in Chaoyang District. (2) The simulation accuracy of the AM-LSTM model is better than that of the traditional LSTM model, and the accuracy of this simulation reaches 22%. (3) The attention weight of the AM-LSTM model indicates that the water level of the second confined aquifer contributes the most to land subsidence. These research findings can provide a reliable model for the prevention and control of land subsidence.关键词：remote sensing;land subsidence;AM-LSTM;simulation and prediction;groundwater level of different layers;attention weight- 1522
- |
- 346
- |
- 0
发布时间：2022-08-08 -
摘要：The rapid and uneven ground subsidence has threatened human production activities, and high-precision subsidence prediction results are of great significance for the precise prevention and control of geological disasters. In order to grasp the evolution law of ground subsidence, a number of prediction studies have been carried out using field observation data or InSAR data. However, due to the existence of spatial heterogeneity, accurate prediction of large-scale ground subsidence is still a challenge.In this study, a spatio-temporal prediction method considering spatial heterogeneity for large-scale ground subsidence STLSTM (Spatio-temporal Long Short-Term Memory) is proposed from a data-driven perspective. First, clustering is used to identify homogenous subregions in geographic space; then, in each subregion, a special Long Short-Term Memory (LSTM) networks are used to capture the nonlinearity features of local locations; Finally, the pre-trained network is used to quantitatively predict the ground subsidence at the future time.In the experimental part, the sentinel-1 image data was used to compare the performance of STLSTM with the other 8 benchmark methods, and the effectiveness of STLSTM was analyzed using spatial statistical indicators. The results show that STLSTM achieves the highest prediction accuracy (71.4%) within 152 secs, and can effectively weaken the effect of spatial heterogeneity on large-scale subsidence prediction tasks.In conclusion, this paper integrates the spatial heterogeneity processing strategy into the deep learning model, and large-scale subsidence prediction is realized with high precision and time efficiency.关键词：remote sensing;ground subsidence;spatio-temporal prediction;heterogeneity;LSTM;InSAR- 974
- |
- 339
- |
- 0
发布时间：2022-08-08 -
摘要：Urban land subsidence is a geological disaster formed by natural and human factors. Cumulative land subsidence easily causes damage to buildings, infrastructure, underground engineering, and other hazards, which seriously threaten the safety of people’s lives and property and cause national economic losses. In the face of urban land subsidence, monitoring, analyzing, and predicting spatiotemporal changes in land subsidence are necessary. The prediction of land subsidence is a crucial step for the early warning of urban infrastructure damage and establishment of a timely remedy.In this study, the Time Series Interferometric Synthetic Aperture Radar (TS-InSAR) technique was utilized to monitor the time series land subsidence at Hong Kong International Airport from 2015 to 2020 by using 152 Sentinel-1A images with an ascending orbit. The local weighted scatter smoothing (Loess) method was used to reduce and smooth the noise in the original data of surface deformation points. Given that the advantages of LSTM correspond to the results of TS-InSAR, on the basis of TS-InSAR data, a stacked LSTM neural network was used to construct a surface deformation prediction model with two LSTM layers, two dense layers, and three dropout layers. The stacked LSTM model was employed to predict the surface deformation of the airport, and its results were compared the predicted results obtained with the true InSAR findings.The average vertical deformation rate of Hong Kong International Airport’s surface for 2015—2020 was -19—5 mm/year. The surface subsidence of the airport gradually increased, and the cumulative subsidence in the vertical direction reached 116 mm in December 2020. The cross-validation of the two time-series analysis methods and the comparison of the monitoring results with the level data showed that the InSAR monitoring results in this study had high accuracy and reliability. A stacked LSTM prediction model was established based on the time-series InSAR monitoring results, and the InSAR observation results were compared with the stacked LSTM prediction results. The root-mean-square error and mean absolute error of the predicted and true values were low, namely, 0.75 and 0.61 mm, respectively, and the correlation coefficient was 0.99. The LSTM prediction model demonstrated good performance at the point level scale and could predict ground subsidence accurately on the basis of TS-InSAR data. The stacked LSTM model was employed to predict the time-series subsidence of Hong Kong International Airport in 2021 by using TS-InSAR deformation data from 2015—2020. The analysis revealed that the long-term prediction process of the stacked LSTM prediction model became invalid after six months. Therefore, the stacked LSTM prediction model is suitable for short-term predictions with a short-term prediction scale of about six months, and the maximum cumulative vertical subsidence of the airport will reach 114 mm in June 2021.In summary, the stacked LSTM prediction model proposed in this study can be used as an effective method to predict surface deformation, and although the LSTM model is only suitable for short-term predictions, its prediction results can be used to assist in decision-making, early warning, and hazard mitigation. In the future, additional data can be incorporated into the LSTM model to accurately determine if the model is suitable for long-term predictions and improve the robustness of the prediction model.关键词：remote sensing;Land deformation;TS-InSAR;Surface deformation prediction;deep learning;LSTM- 1191
- |
- 410
- |
- 1
发布时间：2022-08-08 -
摘要：As a mega city in central China, Zhengzhou is in a period of large-scale metro construction. The problem of ground subsidence along metro lines occurs during metro construction and operation. Thus, monitoring and analysis of ground subsidence along metro lines are important to ensure the safety of metro operation. However, the researches on long-time-series ground subsidence in the Zhengzhou metro network are lacking. Permanent scatterers–interferometric synthetic aperture radar (PS-InSAR) technology overcomes many shortcomings of traditional ground subsidence monitoring methods, such as high cost, limited monitoring range and points, and difficulties in long-term monitoring. Therefore, PS-InSAR technology is used to monitor the subsidence of the Zhengzhou metro network in this study.35 Envisat ASAR images and 44 Sentinel-1 images are employed to obtain the surface deformation information of Zhengzhou City from February 2005 to October 2010 and from July 2015 to May 2019 via PS-InSAR technology. By extracting PS points in a certain range on both sides of the metro lines, the temporal and spatial characteristics of ground subsidence along Zhengzhou metro are subjected to statistical, profile, and overlay analyses. To address the unequal time interval of ground subsidence time series data caused by SAR image discontinuity, an equidistant processing method based on inverse distance interpolation is proposed, and the subsidence of a typical metro station is predicted and analyzed using the Long Short-Term Memory (LSTM) model.Results show that the subsidence sections are mainly concentrated in the east of Lines 1 and 5, the maximum subsidence rate is more than 20 mm/a, and the maximum cumulative subsidence is about 80 mm. The overall deformation trend of Line 1 is similar to a parabola, and the uneven deformation is prominent. The changes in PS points in the time series differ in various regions. The subsidence trough near the Henan Orthopedic Hospital Station of Line 5 is basically symmetrical in space, and the subsidence at the center is expanding yearly. Experiments show that the LSTM model has high prediction accuracy, and the prediction results reveal that the north of the New Archives of Henan Province located in the south of Zhengzhou Sports Center Station will continue to settle at a rate of about 0.5 mm/month in the next two years. Hence, the station and its vicinity must be continuously monitored. This study confirms that PS-InSAR technology can meet the application needs of large-scale urban ground subsidence monitoring, and the results provide a scientific basis for the continuous dynamic monitoring of ground subsidence along Zhengzhou metro network and metro maintenance.关键词：Zhengzhou metro;ground subsidence;PS-InSAR;LSTM;prediction analysis- 1159
- |
- 383
- |
- 2
发布时间：2022-08-08
Regional Land Subsidence
-
摘要：Accurate and rapid regional-scale crop yield estimation can provide effective data support for the formulation of national food security policies. Compared with complex mechanism models, sampling statistical surveys and empirical models based on multi-source data have better reliability and operability for county-level or city-level yield estimation. Previous studies have proposed many factors related to winter wheat yield, but systematic research on the selection and analysis of multi-source factors is lacking. On the basis of remote sensing, meteorological, and statistical data, this study systematically explored the influence of key-phase growth-environment-landscape features on winter wheat yield estimation at the county-level and determined the best time phase and characteristic parameters.The considered features included crop condition, environmental forcing (e.g., precipitation, light, and temperature), and farmland landscape. The key phases were the key periods of winter wheat yield formation (P1—P5), which were extracted from the NDVI curve of the crop growth process. Random Forest (RF) regression models were developed using different combinations of phases and features to simulate statistical wheat yield data and evaluate the importance of different combinations via an accuracy assessment. The performance of the models built from each layer combination was compared using the Mean Relative Error (MRE), Root Mean-Squared Error (RMSE), Normalized Root Mean-Squared Error (NRMSE), and coefficient of determination (R2). Data on years 2014—2017 were used to build the models, and 2018 data were utilized for validation.Results showed that P2, P3, and P4 resulted in higher accuracy than P1 and P5 in terms of single phases. The model accuracy using multi-phase features was higher than that of using single phases, and the combination of P2 and P4 was the best. Among all the features, crop growth features had the greatest impact on yield estimation accuracy, whereas the addition of environmental forcing factors (e.g., water, light, and temperature) did not significantly improve the accuracy. The addition of farmland landscape features could effectively improve the accuracy of yield estimation. Moreover, five important features (PROP, NDVI_P2, B2_P2, ED, and B1_P4) were selected, and a yield estimation model was established to obtain the county-level yield of winter wheat in Hebei Province. The MRE of wheat yield estimation at the county level in 2018 was as low as 2.85%, and the RMSE, NRMSE, and R2 were 253.25 kg/ha, 4.09%, and 0.83, respectively.Conclusion Multi-phase performance is better than single-phase performance. Combining crop growth features with farmland landscape features (RMSE of 247.79 kg/ha) provides more accurate estimates than using crop growth features alone (RMSE of 295.95 kg/ha). Furthermore, the RF model produces good yield estimation results. This study provides insights into and new methods for nationwide estimation of winter wheat yield at the county level.关键词：remote sensing;yield estimation;winter wheat;statistical data;NDVI;Random Forest;Hebei Province- 1067
- |
- 292
- |
- 2
发布时间：2022-08-08 -
摘要：Timely and accurate estimation of the spatial distribution of cropland is critical for agricultural production management, yield estimation, and planting structure adjustment. Previous studies on cropland mapping mostly focused on using moderate-/low-spatial-resolution images or single-phase high-spatial-resolution images, in which croplands in regions with fragmented landscapes and complex crop planting patterns are challenging to extract. The multi-source Gaofen (GF) satellites launched by China can provide images with high spatiotemporal resolution, thus presenting great potential for fine-scale cropland mapping with high accuracy. This study utilized GF-1, GF-2, and GF-6 satellites to explore a high-accuracy cropland mapping method at metric spatial resolution. Specifically, Cropland Extraction UNet (CEUNet) was developed based on the structure of UNet by integrating multi-temporal information from GF-1/6 and spatial details from GF-2 to fully exploit the spatial and temporal characteristics of cropland.To make full use of the details provided by high-spatial-resolution images, CEUNet adopted the same encoder structure as UNet. It consisted of the repeated application of two 3×3 convolutions (unpadded convolutions), each followed by a 2×2 max pooling operation with stride of 2 for down sampling. The input image size was gradually reduced at each down sampling step, and feature maps of each size were concatenated in the corresponding up-sample layer to extract the hierarchy of spatial information. Meanwhile, time-series feature maps of images with moderate to high spatial resolution were extracted by two consecutive 3×3 convolutional layers and a 1×1 convolutional layer. Then, the time-series and spatial feature maps were integrated via element-wise addition before they were sent to the decoder for pixel-wise classification.Evaluation results from randomly selected sample points showed that CEUNet achieved good performance with an overall accuracy of 92.92% over the whole of Qianjiang City, Hubei Province. The CEUNet-extracted cropland at the meter-level resolution was used to perform a wall-to-wall pixel comparison with the cropland extracted via semantic segmentation based on UNet by using multi-source remote sensing images with different resolutions (UNet_m). Semantic segmentation based on UNet using single-phase high-resolution images (UNet_s), object-based random forest classification (OBIA), and pixel-based random forest classification (RF) were employed to extract the results of cultivated land for comparison. The accuracy of cropland extraction by CEUNet was higher than that by others (the average F1 score was improved by about 0.04, 0.11, 0.21, and 0.21), indicating the effectiveness of the proposed approach for diverse agriculture landscapes. The F1 score of CEUNet was improved by about 0.09, 0.26, 0.27, and 0.27 over regions with high fragmentation and complex landscapes.关键词：cropland extraction;multi-source remote sensing images;CNN;GF satellites- 1629
- |
- 686
- |
- 3
发布时间：2022-08-08 -
摘要：Accurate farmland area identification is the basis of crop yield estimation and an important indicator in food security assessment. As an important data source for farmland identification, remote sensing data can provide dynamic and fast observation results for classification. GF-5, which is the only hyperspectral satellite in the China High-resolution Earth Observation System, has great research and application potential in farmland identification. However, the dimensionality curse caused by the redundant bands in hyperspectral data seriously affects the calculation speed and classification accuracy of models. To solve this problem, this research proposes a hybrid feature selection algorithm for farmland identification. First, on the basis of the feature importance provided by the feature selection algorithm, the feature dimension is gradually reduced from 295 to 5 with a step length of 10. The overall accuracy of the classification results corresponding to each feature dimension is recorded. Second, the turning point (a dimension number whose corresponding overall accuracy hardly decreases when the input variable number is smaller than it) is determined based on the overall accuracy, and the corresponding variables are adopted as the feature subset. Lastly, the Sequential Backward Selection (SBS) method is used to search for the best subset.Three feature selection algorithms (i.e., Random Forest (RF), Multi-Information (MI), and L1 regularization (L1)) and three classification algorithms (RF, Support Vector Machine (SVM), and K-Nearest Neighbor (KNN)) are examined. Results indicate that the autocorrelations of the three subsets differ significantly. Most of the bands selected by the MI method are continuous and concentrated in the blue and shortwave infrared range. Therefore, the extremely high autocorrelation that exists in this subset has a negative effect on classification accuracy. By contrast, the correlation between bands in the RF and L1 feature subsets is relatively weak. However, the two feature sets still result in different classification accuracy. According to the variable distribution, many red-edge and near-infrared bands are contained in the L1 feature subset. These bands demonstrate better ability to distinguish farmland, forest, and soil than the blue and red bands selected by the RF algorithm. The classification algorithms also have different capacities. In the high-dimensional space, the SVM algorithm exhibits high robustness to noise, resulting in high accuracy. However, when the dimension decreases to a critical value, the accuracy of SVM decreases sharply. By contrast, although RF is not as robust as SVM in the high-dimensional space, it has excellent generalization ability in the low-dimensional space. Compared with the subsets obtained after the first dimensionality reduction process, the optimal feature subsets obtained by SBS searching improve the classification accuracy of each model.The L1-SVM-SBS model with a 23-dimensional input achieves the highest overall classification accuracy (94.64%) and cropland recall rate (95.83%). This study provides a new method of farmland identification using hyperspectral data. By selecting numerous representative and informative bands, this method not only improves farmland classification accuracy, but can also be used as a reference for other classification problems involving hyperspectral remote sensing.关键词：cropland identification;GF-5;feature selection;hyperspectral remote sensing;L1 regularization;sequential backward selection- 1206
- |
- 384
- |
- 0
发布时间：2022-08-08 -
摘要：Driven by big data, deep learning has been widely and successfully applied in many fields, such as computer vision and speech recognition. With the increase in network depth, deep learning models can determine the rules and layers of images and obtain high classification accuracy, so they have become a research hotspot in remote sensing image interpretation. As data-driven algorithms, deep learning models need a large number of labeled samples for training to ensure that the trained model can learn accurate and comprehensive sample features, and they exhibit good classification performance. However, although the development and maturity of remote sensing technology provide abundant remote sensing image sources for deep learning models, the application of deep learning technology in remote sensing is limited by expensive manually labeled samples,. In practical crop classification applications, the quantity and quality of existing ground truth samples are often insufficient to train a classification model with high performance.This study proposes a crop classification strategy for the deep learning model based on weak samples to verify the applicability of the deep learning model with weak samples.GF-1 was used as the data source, and the SVM classifier was used to classify three types of rice, corn, and other ground objects in Liaoning Province at the county level. The results were used as the training label samples of the deep learning model. This process included sub-county SVM classification, manual post-classification processing, cropland masking, and other operations. This human-computer interaction was chosen to ensure the accuracy of the results. In this study, samples with non-100% accuracy were labeled as weak samples. Then, a Deep Convolutional Neural Network (DCNN) model was used to train the weak samples and obtain the spatial distribution of rice and corn in Liaoning Province.Results showed that OA reached 0.90, and the F1 scores of rice and corn were 0.81 and 0.90, respectively. The spatial consistency with the SVM results was 0.90. The model showed good robustness under the different topography and landform types of the agricultural landscape with a median OA that was greater than 0.93. It overcame the influence of topography in the study area to a certain extent through subregion analysis. In the agricultural landscape with a complex planting structure, the proposed method still maintained a certain accuracy in crop classification. Subsequently, noise experiments were designed to analyze the influence of SVM label noise on model classification. The corn distribution in the original SVM training label was expanded from 1 to 40 times to obtain new labels, which were then used to train the DCNN model and predict the testing data. When the model was within five times the sample noise, that is, the sample maximum error area ratio was not more than 0.36, the model was robust to a certain extent, and the results could be maintained within a reliable accuracy range (OA remained to be greater than 0.86).In conclusion, this study verified that crop classification results obtained with the deep learning model whose training labels are based on traditional classification methods can achieve high recognition accuracy good robustness under different topography and landform types of agricultural landscapes and the feasibility of using traditional classification results as weak samples. The experiment on increasing noise in the weak samples showed that weak samples can be used to train DCNN as long as their identification accuracy is guaranteed, that is, the maximum error area ratio of samples is not more than 0.36. This approach further reduces the threshold of obtaining labeled samples via deep learning models. It makes up for the limitation of the deep learning model, which is highly dependent on a large number of manually labeled samples, and provides a new approach for large-area remote sensing crop classification.关键词：Weak samples;Deep Convolutional Neural Networks (DCNN);deep learning;GF-1;Crop remote sensing classification- 1149
- |
- 406
- |
- 4
发布时间：2022-08-08 -
摘要：Corn and soybean are two major crops maintaining food security, and thus the timely and accurate monitoring of their planting areas are of great importance to forecasting their production and market prices. The objectives of this study were to use a remote-sensing technology in exploring indicative features that can effectively identify corn and soybean in their middle and late growth seasons and provide technical support for the broad geographical application of corn and soybean mapping. This study can facilitate the early release of corn and soybean planting acreages for policy makers. In this study, two typical planting areas of corn and soybean in the provinces of Heilongjiang and Anhui were selected. GaoFen-1 satellite images with 30 m spatial resolution were acquired in the middle and latter growth stages and used as data sources for calculating various vegetation indices and textural features. Then, a feature optimization method was used in evaluating the relative importance scores of input features, and optimal feature combinations for identifying corn and soybean were determined. The random forest classification algorithm was used in analyzing the relationship between the number of input features and classification accuracy, and then the best feature groups in different experimental areas were identified. Finally, according to similarities and differences among the selected features in different regions, the indicative features for mapping corn and soybean in the middle and latter stages were established. The validity and stability were confirmed using our experimental designs. The following results were obtained: (1) indicative remote sensing features for efficiently identifying corn and soybean in their middle and late growing seasons were identified; (2) the classification performance of the indicative features of corn and soybean in both experimental areas was approximately 10% higher than that when original spectral band combinations were used. In different planting areas, high classification accuracy was obtained using the indicative features of corn and soybean as the optimal features selected in individual local area. Our selected indicative features for soybean and corn mapping were found stable, effective, and useful for large areas of implementation. These features included Ratio Vegetation Index (RVI), Difference Vegetation Index (DVI), Conversion Vegetation Index (TVI), improved chlorophyll absorption ratio index (MCARI), and the second moment and entropy in Gray Level Co-occurrence Matrix (GLCM).关键词：remote sensing;corn;soybean;remote sensing identification;satellite feature;classification;GF-1- 1433
- |
- 312
- |
- 1
发布时间：2022-08-08 -
摘要：Rice-crayfish co-culture is a kind of comprehensive ecological agriculture pattern. Rice-crayfish co-culture has expanded rapidly in China in the past decade due to its outstanding ecological and economic benefits. The accurate spatial distribution information of this newly emerging agricultural pattern is crucial for growth monitoring, yield estimation, and water resource management. However, most studies have focused on field-level research on the farmland ecosystem, and rice-crayfish mapping at regional or larger scales has received less attention.In this study, Qianjiang City in Hubei Province, known as the “hometown of crayfish” was selected as the test area. The cloud computing approach was used for all available Landsat 7/8 and Sentinel-2 imagery in 2019 with the Google Earth Engine (GEE) platform. By analyzing farming characteristics and the spectral curves of rice-crayfish fields, we identified the crucial phenology windows and classification features (i.e., flooding and vegetation signals) for rice-crayfish mapping. On the basis of the key phenological characteristics and associated frequency thresholds derived from field samples, we developed a rule-based algorithm for rice-crayfish mapping and generated the rice-crayfish map of Qianjiang City in 2019. To further evaluate the potential of our proposed method, we compared it with the random forest method and a method based on seasonal differences of water bodies.The spectral analysis of time series images showed that the unique phenological characteristics of rice-crayfish co-culture were flooding signal (LSWI > NDVI or LSWI > EVI) in phenology window 1 (from January 1 to April 30), vegetation signal (NDVI > LSWI or EVI > LSWI) in phenology window 2 (from July 15 to September 30), and flooding signal in phenology window 3 (from November 10 to December 31). With the mapping results, the total area for rice-crayfish planting in Qianjiang City in 2019 was estimated to 575.58 km2, and rice-crayfish plots were mainly distributed in the southwest. The producer’s accuracy of the classification result was 90.74%, the user’s accuracy was 94.69%, and the overall accuracy was 95.23%. The method based on phenology window features had fewer commission errors compared with the random forest method and fewer omission errors compared with the method based on water body seasonal differences. Among the three methods, the proposed method presented the highest overall classification accuracy.The rice-crayfish mapping method based on phenology windows, flooding signal, and vegetation signal showed high separability. The method based on phenology windows can be easily generalized to other regions and other images because of its strong physical interpretation for rice-crayfish. On the one hand, this method has relatively low dependence on training samples. On the other hand, as long as the key phenology window can be obtained, other medium-high resolution images, such as GF-1 and GF-6, can achieve high-accuracy mapping results for rice-crayfish. Therefore, the method based on phenology windows can be effectively extended to large areas and long time series. It can provide essential information for rice production management and decision-making in the crayfish industry.关键词：remote sensing;rice-crayfish;crop extraction;Google Earth Engine;phenology window;Landsat;Sentinel-2- 899
- |
- 301
- |
- 3
发布时间：2022-08-08 -
摘要：Crop planting area and its spatiotemporal distribution information are crucial for agricultural management, structural adjustment of the planting industry, national food security, and other fields. Remote sensing can extract crop planting information quickly. However, a large amount of ground survey information, expert knowledge, and manual correction operation after classification are needed to meet actual production needs. Using remote sensing and existing geographic information data for intelligent information extraction is the future development trend. The purpose of this study is to investigate the classification representation method of the deep learning model on the basis of the constraints of field parcel data. Multi-source heterogeneous data are linked by constructing representation. The method can identify the spectral differences of different crops in different phases and classify the crop types.The study area is in the eighth division of Xinjiang Production and Construction Corp. The platform of Google Earth is used to obtain all Landsat 8 images during the crop growth period of the research area in 2019. The image data and their quality assessment band on GEE are utilized to determine the statistical or estimated values of phases in the study area. Then, the average reflectivity of each band of the extracted block level is arranged in accordance with the time-phase sequence and wavelength, and a plot representation of the spectrum and time-phase two-dimensional properties is constructed. The construction of land plot representation realizes the connection of geographic information and remote sensing data and makes the application of deep learning in crop remote sensing classification possible. The completed plot representation is used to construct and train a Convolution Neural Network (CNN) model.A step-by-step optimization process is implemented to search for the best combination of super parameters and various types of layers. Construction of a relatively optimal CNN model is obtained, and classification of the constructed feature map is carried out based on this model to complete the classification of the research area in 2019. We obtain a crop distribution map whose resolution is much higher than that of the remote sensing image. Moreover, the overall accuracy of the CNN model reaches 93.04%, and the kappa coefficient is 91.09%. The results of nine kinds of crop classification are good. After thousands of rounds of data learning, the proposed method exhibits lower classification error fluctuation and higher stability than other machine learning algorithms.The research object is plot representation, which is the abstract expression of plot planting information. It can be used as the standard input of the deep learning model. Through the construction of plot representation, crop classification can be indirectly identified by remote sensing. The proposed method has outstanding application potential in land feature representation and crop classification and should be regarded as an optimal method for multi-temporal image classification tasks based on field parcel data. The method can be used as a reference in the application of deep learning in remote sensing.关键词：remote sensing;crop classification;CNN;Feature map;Plot representation;multi-temporal image;Per-field- 805
- |
- 272
- |
- 2
发布时间：2022-08-08
Agricultural Remote Sensing
-
摘要：Icebergs are formed either by the calving of the seaward margins of floating glacier tongues or ice shelves or by the fragmentation of existing icebergs. Mass loss caused by iceberg calving represents up to half of the total loss of mass from the Antarctic ice shelves. As they melt and drift with the ocean currents, icebergs provide a significant source of freshwater input to the surface layer of the ocean, enough to affect the stability of stratification in the upper ocean. Any increased freshwater in iceberg discharge, or ice shelf melt, likely has a major impact on the ocean circulation by delivering sufficient freshwater, which will play a critical role in many geophysical and biological processes. The iceberg freeboard is an important geometric parameter for measuring the thickness of an iceberg and estimating its volume. Because of the paucity of high-precision measurements on iceberg thickness, large uncertainties exist on the iceberg’s ice volume estimation of the Southern Ocean. The freeboard of large icebergs has been successfully extracted using the altimetry method. These large uncertainties exist, however, because of the insufficient altimeter coverage of various icebergs. On the basis of the fact that an iceberg can cast an elongated shadow on the surface of sea ice in winter, this study proposes a method to measure the iceberg freeboard by using shadow length and the predefined or estimated solar elevation angle. Three Landsat 8 panchromatic images with center solar elevation angles of 5.43°, 7.49°, and 11.01° on August 29, September 7, and September 16 in 2016, respectively, are selected to test our method. The shadow lengths of five isolated tabular icebergs are automatically extracted to calculate the freeboard height. For an accuracy assessment, we perform cross-validation on the matching points at different times. Results show that the measurement error of shadow length is less than one pixel. When the sun elevation angle is lower than 11.01°, the root-mean-square error (RMSE) of the iceberg freeboard from the 15 m panchromatic image is less than 2.0 m, and the mean absolute error (MAE) is less than 1.5 m. The experiment shows that under the angle of low solar elevation in winter, the Landsat 8 15 m panchromatic images can be used for high-precision measurement of the iceberg freeboard and has the potential to measure the Antarctic iceberg freeboard at a large scale.关键词：remote sensing;Antarctic;icebergs;freeboard;shadow altimetry;Landsat 8- 708
- |
- 235
- |
- 0
发布时间：2022-08-08 -
摘要：High-density urban cities contain numerous similar buildings positioned in close proximity. Building detection from high-spatial-resolution remote sensing imagery in such scenes remains a challenge in computer vision and remote sensing urban applications. The integration of traditional segmentation algorithms and a novel neural network is an effective approach for such challenging settings. Inspired by the recent success of deep-learning-based edge detection, a new building detection method aiming at accurate boundaries is proposed. In accordance with the characteristics of buildings and their border, this study improves the network structure and integrates the network with bottom-up watershed segmentation to improve boundary precision and classification accuracy.First, two auxiliary labels, namely, the building boundary and parting line, are derived from the original dataset through data preprocessing. Second, the newly proposed building detection frame called ICT-Net is improved by modifying its structure and loss function in accordance with the two auxiliary labels to obtain the probability of three classes. Lastly, a post-process integrating watershed segmentation with gradient-boosted regression trees is employed to achieve high-accuracy building detection. Specifically, a probability feature map is generated by merging the probability of three classes. Watershed segmentation with building marker thresholds is applied to obtain building instances from the probability feature map. Then, the building probability of each building instance predicted by gradient-boosted regression trees is used to select building instances, resulting in building detection results. Parameter selection is also implemented.The performance of the proposed method is validated on the INRIA dataset, which provides aerial orthorectified color imagery with a spatial resolution of 0.3 m and with corresponding ground truth labels for two semantic classes: building and not building. Experimental results suggest that data preprocessing and the application of boundary loss can obtain an improvement of 1% in terms of the Intersection over Union (IOU) of building detection. The post-process can take full advantage of probability information from the network, thereby effectively optimizing the building boundary. The post-process brings an improvement of 10.5% in terms of building instance recall compared with the results of the neural network. Our study achieves a building instance recall rate that is 22.9% higher than that of the original ICT-Net.A novel building detection method based on a boundary-regulated network and watershed segmentation is proposed in this study. Experimental results reveal the advantages of the enhanced-boundary-oriented data preprocessing and modified neural network and demonstrate that the proposed method can further improve prediction accuracy on a network basis. However, the excellent performance of the proposed method largely depends on parameter selection, and further improvements should be made in the future.关键词：remote sensing;building detection;deep neural network;boundary loss;watershed segmentation;instance segmentation- 896
- |
- 266
- |
- 2
发布时间：2022-08-08 -
摘要：As the process of urbanization accelerates, changes in the physical structure and attributes of the natural surface have increased the impervious surface of cities, resulting in drastic changes in the types of urban land cover, which greatly affect environmental quality and the ecological cycle. Therefore, exploring the spatial structure of urban impervious surfaces crucial to urban ecological environment protection and urban green space planning.The main urban areas of typical Xinjiang cities (Urumqi, Kashgar, Hami, and Karamay) are selected as the study area in this work. First, the L2 norm of Sentinel-2A/B images is normalized to weaken the urban bare soil interference information in arid areas. Combined with the Enhanced Normalized Difference Impervious Surface Index (ENDISI), the maximum between-class variance method (OTSU) is used to adaptively determine the threshold, and the impervious surfaces of typical cities in Xinjiang in 2017 and 2019 are extracted. Second, the spatial differences of the ENDISI values in each direction of impervious surfaces in the study area are analyzed with the profile line method, and the box dimension method is introduced to dissect the spatial structure characteristics of impervious surfaces in different time phases and reveal the law of urban spatial structure changes.Results show that the combination of L2 norm normalization treatment and ENDISI can effectively highlight the difference between impervious and non-impervious surfaces. The threshold value determined by OTSU adaptively can distinguish impervious surfaces well, and the overall accuracy of the impervious surface extraction results is 86.60% with a kappa coefficient of 0.73 as verified by an example (2019 Urumqi main urban area). With the junction area of a city and bare soil as the research object, the classification effects of ENDISI, NDBI, and UI methods are compared, and the accuracy is verified using the visual interpretation method. The results show that ENDISI’s extraction effect is the best, and its extraction accuracy for urban contours and bare soil areas is the highest (83.18%). NDBI’s extraction effect is relatively poor (78.38%), and UI’s extraction accuracy is the lowest (60.18%).The analysis of the spatial differences of the impervious surfaces reveals that the ENDISI values for northern Urumqi, central–northern Kashgar, and central-eastern and northern Hami increase significantly from the perspective of profile lines, but the ENDISI values for northern Karamay and central-western regions increase only slightly. The analysis of the box dimension of the impervious surface shows that the box dimension values of typical cities in Xinjiang and the urban structure complexity are increasing. Among these cities, Hami and Urumqi have the highest and lowest box dimension values, respectively; Hami has the largest variation in box dimension values, whereas Karamay has the smallest. In this study, a new method of extracting urban impervious surfaces in arid areas is proposed. The method reveals the change law of impervious surfaces of cities in various directions, provides scientific guidance for the connotative development of typical cities in Xinjiang, and serves as a theoretical basis for the protection of urban ecological environments in arid areas.关键词：remote sensing;urban impervious surface;Sentinel-2A/B;ENDISI;OTSU;The Box-counting Dimension;Xinjiang Uygur Autonomous Region of China- 944
- |
- 266
- |
- 2
发布时间：2022-08-08
Remote Sensing Technology and Method
- Postal code：100190
- Tel：010-58887052 Email：nrsb@aircas.ac.cn
- Technical support is provided by Beijing Founder electronics co., LTD 京ICP备20021838号-9
京公网安备11010802024621 - It is recommended to read the content of this site in Chrome&IE9+. Please switch to extreme mode in browser 360.
- Cookies We use cookies to help provide and enhance our service and tailor content. By continuing, you agree to the use of cookies.
0