Accurate extraction of crop cultivated area and spatial distribution is essential for food security. Crop classification methods based on machine learning and deep learning and remotely sensed time-series data are widely utilized to detect crop planting area. However, few studies assess the effectiveness of machine learning and deep learning algorithm integrated time-series satellite data for identifying multiple crop type classification over a large-scale region. Hence, this study aims to evaluate the effectiveness of machine learning and deep learning models in crop classification and provide a framework for large-scale multiple crop type classification based on time-series of satellite data. The time-series of the Normalized Difference Vegetation Index (NDVI), Enhanced Vegetation Index (EVI), and NaE (combined NDVI and EVI) were adopted as input features, and four widely used machine learning models, including Support Vector Machine (SVM), Random Forest (RF), K-Nearest Neighbor (KNN), and their integration (Stacking) were compared to examine the performance of multiple crop types (i.e., wheat, corn, wheat-corn, early rice, and early rice-late rice) classification in the North China Plain. The performance of two types of deep learning classifiers: the One-dimensional Convolutional Neural Network (Conv1D) and Long Short-Term Memory Networks (LSTM), were also tested. The results showed that the NaE feature performed best among three input features, and the Stacking model produced the highest accuracy (77.12%) compared to other algorithms.

Satellite-borne multispectral data are suitable for regional-scale grassland community classification owing to comprehensive coverage. However, the spectral similarity of different communities makes it challenging to distinguish them based on a single multispectral data. To address this issue, we proposed a support vector machine (SVM)–based method integrating multispectral data, two-band enhanced vegetation index (EVI2) time-series, and phenological features extracted from Chinese GaoFen (GF)-1/6 satellite with (16 $\mathrm{m}$) spatial and (2 d) temporal resolution. To obtain cloud-free images, the Enhanced Spatial and Temporal Adaptive Reflectance Fusion Model (ESTARFM) algorithm was employed in this study. By using the algorithm on the coarse cloudless images at the same or similar time as the fine images with cloud cover, the cloudless fine images were obtained, and the cloudless EVI2 time-series and phenological features were generated. The developed method was applied to identify grassland communities in Ordos, China. The results show that the Caragana pumila Pojark, Caragana davazamcii Sanchir and Salix schwerinii E. L. Wolf grassland, the Potaninia mongolica Maxim, Ammopiptanthus mongolicus S. H. Cheng and Tetraena mongolica Maxim grassland, the Caryopteris mongholica Bunge and Artemisia ordosica Krasch grassland, the Calligonum mongolicum Turcz grassland, and the Stipa breviflora Griseb and Stipa bungeana Trin grassland are distinguished with an overall accuracy of 87.25%. The results highlight that, compared to multispectral data only, the addition of EVI2 time-series and phenological features improves the classification accuracy by 9.63% and 14.7%, respectively, and even by 27.36% when these two features are combined together, and indicate the advantage of the fine images in this study, compared to 500 $\mathrm{m}$ moderate-resolution imaging spectroradiometer (MODIS) data, which are commonly used for grassland classification at regional scale, while using 16 $\mathrm{m}$ GF data suggests a 23.96% increase in classification accuracy with the same extracted features. This study indicates that the proposed method is suitable for regional-scale grassland community classification.

Evaluating the potential productivity of the terrestrial ecosystem is extremely important to ascertain the threshold of vegetation productivity, to maximize the utilization of regional climate resources, carbon sequestration and to mitigate climate warming caused by rising CO2 concentrations. However, most previous studies neglected the optimum state of natural vegetation without human intervention and regional change trend of vegetation under future climate change. In this study, variations in spatio-temporal distributions of climate potential productivity (CPP) over China from 1980 to 2018 are analyzed with the synthetic estimating model. A comprehensive regionalization method (Principal components analysis, PCA) based on standardized precipitation evapotranspiration index (SPEI), and statistical analysis methods are adopted to assess CPP and its response to the climate change in different regions of China. The results demonstrate that the global temperature rising and precipitation decreasing have obvious effects on the productivity of terrestrial ecosystem and its spatio-temporal distribution in different sub-regions and ecosystems. Among them, precipitation is the dominant factor, and temperature significantly affects some regions such as Tibetan Plateau (TP) and Northeast China (NE) with high-altitude or high-latitude. The optimum temperature for the CPP in Xinjiang (XJ) region and Northwest China (NW) is 7.5℃ and 8℃, respectively. With regards to the ecosystems, the CPP of grassland shows complex trends in XJ, Southwest China (SW), NE, and TP; especially in XJ (NE), the CPP shows a decreasing (an increasing) trend when the temperature is more than 7.5℃ (0℃). Linear correlations occur between farmland CPP and temperature in each sub-region except for XJ. The same situation also exists at forest CPP, especially in TP, NE and NC regions. However, under the temperature increasing and precipitation decreasing, there are slight adverse impacts on the CPP of vegetation at the national scale, indicating that drier and warmer climate are detrimental for vegetation growth.

Remote sensing (RS) is a convenient technology to estimate the regional cultivation areas of crops. However, the accurate estimation of maize areas using RS over a broad region is a significant challenge due to the large phenology differences and insufficient prior knowledge in space. To address this issue, a new method was developed in this work. In this method, the correlation (r ) and root mean standard error (RMSE) between the time-series moderate resolution imaging spectroradiometer enhanced vegetation index (MODIS EVI) and the standard EVI curve of maize from a reference area are computed. Pixels with a high value of r and a low value of RMSE were identified as maize areas. The phenology information observed at agro-meteorological stations was also used to recognize maize pixels from the pixel-level phenology derived from time-series MODIS EVI. The proposed method provides an accurate characterization of the phenology differences over the study area by making use of the planting and maturity dates only. In addition, the few location-dependent parameters make the recognition of maize planting areas over large regions easier than previous studies. The proposed method was implemented over the Northeast China Plain (NECP) and North China Plain (NCP). The derived results were compared with official statistical results, and a close agreement was observed. At the city level, the satellite-derived estimates agreed well with the statistics with the R ² (RMSE) of 0.86(110.97 k hm ² ) in the NECP and 0.76(68.74 k hm ² ) in the NCP. At the county level, the R ² (RMSE) is 0.82(25.47 k hm ² ) in the NECP and 0.75(5.93 k hm ² ) in the NCP. At both temporal levels, the R ² (RMSE) results obtained in this work are higher(lower) than those published in other studies. The obtained results indicate that the proposed method is effective in maize area estimation over broad regions.

Recently, convolutional neural networks (CNNs) have been widely used in hyperspectral image (HSI) classification with appreciable performance. However, the current CNN-based HSI classification methods have limitations in exploiting the multiscale features and extracting sufficiently discriminative features, and usually, adopted dimensionality reduction method such as principal component analysis (PCA) leads to some or all of the physical information of the original band may be lost. To address the above problems, in this letter, we propose an adaptive multiscale feature attention network (AMFAN) for HSI classification. First, we use a band selection algorithm to perform data dimensionality reduction, which helps maintain the original characteristics of the image. Second, different from existing multiscale feature extraction methods that give features of different scales the same degree of importance, we propose an adaptive multiscale feature residual module (AMFRM) to give multiscale features different importance. Finally, due to the input of the HSI classification model based on deep learning (DL) being the patch cube, the only available initial information is the category of the center pixel. However, the patch often contains pixels different from the center pixel category, and existing attention mechanisms do not consider the impact of such pixels on the HSI classification, so we design a novel position attention module (PAM) to calculate the similarity between the center (target) pixel and surrounding pixels and then pay more attention to the pixels with high similarity to the center pixel. Besides, we also use a spectral attention module (SAM) to obtain more discriminative spectral features. Experimental results show that the proposed AMFAN effectively improves the classification accuracy and outperforms the state-of-the-art CNNs.

The accurate and timely spatial distribution information of various crop types is vital for food security. In this study, the 2-band enhanced vegetation index (EVI2) data from Moderate Resolution Imaging Spectroradiometer (MODIS) were combined with sparse representation approach to identify the distribution of various crop types in Shandong Province, China. Three groups of input variables derived from EVI2 including annual time series EVI2 (TS), harmonic features (HF), and combined vector formed by harmonic features and texture features (HFT) were used. The online dictionary learning and orthogonal matching pursuit algorithms were applied to generate the dictionary and solve the sparse coefficients of the identified samples, respectively. Then, the label of the identified samples can be obtained according to the minimum residuals between the dictionary and the sparse coefficients of the identified samples. At the provincial level, the validation based on the statistical data showed that three groups of input variables presented lower than ±25% at relative errors, and input variables of HFT performed better than the other two. At the municipal level, the results achieved by using input variables of HFT also agreed well with the statistics with the coefficient of determination R ² > 0.85 for wheat and maize, as well as R ² > 0.71 for peanut and cotton during 2014-2016. These results demonstrate that the combination of the input variables of HFT derived from time series MODIS EVI2 data and sparse representation approach can be used for crop identification in the study area.