Timely and accurate agricultural information is essential for food security assessment and agricultural management. Synthetic aperture radar (SAR) systems are increasingly available in crop mapping, as they provide all-weather imagery. In particular, the Sentinel-1 sensor provides dense time-series data, thus offering a unique opportunity for crop mapping. However, in most studies, the Sentinel-1 complex backscatter coefficient was used directly which limits the potential of the Sentinel-1 in crop mapping. Meanwhile, most of the existing methods may not be tailored for the task of crop classification in time-series polarimetric SAR data. To solve the above problem, we present a novel deep learning strategy in this research. To be specific, we collected Sentinel-1 time-series data in two study areas. The Sentinel-1 image covariance matrix is used as an input to maintain the integrity of polarimetric information. Then, a depthwise separable convolution recurrent neural network (DSCRNN) architecture is proposed to characterize crop types from multiple perspectives and achieve better classification results. The experimental results indicate that the proposed method achieves better accuracy in complex agricultural areas than other classical methods. Additionally, the variable importance provided by the random forest (RF) illustrated that the covariance vector has a far greater influence than the backscatter coefficient. Consequently, the strategy proposed in this research is effective and promising for crop mapping.

Change detection by comparing two bitemporal images is one of the most challenging tasks in remote sensing. At present, most related studies focus on change area detection while neglecting multiple change type identification. In this letter, an attention gates generative adversarial adaptation network (AG-GAAN) is proposed on multiple change detection. The AG-GAAN has the following contributions: 1) this method can automatically detect multiple changes; 2) it includes attention gates mechanism for spatial constraint and accelerates change area identification with finer contours; and 3) the domain similarity loss is introduced to improve the discriminability of the model so that the model can map out real changes more accurately. To demonstrate the robustness of this approach, we used the Google Earth data sets that include seasonal variations for change detection and understanding. The experimental results demonstrated that the proposed method can accurately detect the multiple change types from bitemporal imagery.

Hyperspectral image analysis plays an important role in agriculture, mineral industry, and for military purposes. However, it is quite challenging when classifying high-dimensional hyperspectral data with few labeled samples. Currently, generative adversarial networks (GANs) have been widely used for sample generation, but it is difficult to acquire high-quality samples with unwanted noises and uncontrolled divergences. To generate high-quality hyperspectral samples, a self-attention generative adversarial adaptation network (SaGAAN) is proposed in this work. It aims to increase the number and quality of training samples to avoid the impact of over-fitting. Compared to the traditional GANs, the proposed method has two contributions: (1) it includes a domain adaptation term to constrain generated samples to be more realistic to the original ones; and (2) it uses the self-attention mechanism to capture the long-range dependencies across the spectral bands and further improve the quality of generated samples. To demonstrate the effectiveness of the proposed SaGAAN, we tested it on two well-known hyperspectral datasets: Pavia University and Indian Pines. The experiment results illustrate that the proposed method can greatly improve the classification accuracy, even with a small number of initial labeled samples.

Accurate information on cropland changes is critical for food production and security, sustainable cropland management, and global change studies. The common change detection methods bi-temporal based, using remotely sensed imagery easily generate pseudo changes due to phenological or seasonal differences. Cropland exhibits a distinctive phenological trajectory that has strong periodic characteristics and seasonal paths. This paper proposes the use of phenological trajectory similarity to search for the overall changes between two time-series images instead of single change events between two dates of imagery. Due to the complex spectral–temporal characteristic of cropland, a phenological trajectory was constructed using a multi-harmonic model for capturing intra-annual variations. Then, phenological trajectory similarity was measured using coefficient vector difference (CVD), and used for detecting change/no-change areas when considering both the amplitude and phase difference. Finally, instead of the traditional classification method based on original images, we used the coefficient ratio vector (CRV) as the input for change type discrimination. The distance between the coefficient ratio vector (CRV) of the change pixel and of the reference change type was calculated to identify the exactly changed types. The performance of this proposed approach was tested using two sets of Landsat time-series images from 2010 and 2015. Moreover, the change area detection results of three other methods, namely, the continuous change detection and classification (CCDC), change vector analysis (CVA), and post-classification comparison (PCC), were also calculated for comparison and analysis. The results indicated that the proposed approach acquired the highest accuracy with an overall accuracy of 98.58% and a kappa coefficient of 0.82, which demonstrated that the method provides the capacity to detect real changes and estimate pseudo changes caused by season differences.

Spectral and NDVI values have been used to calculate the change magnitudes of land cover, but may result in many pseudo-changes because of inter-class variance. Recently, the shape information of spectral or NDVI curves such as direction, angle, gradient, or other mathematical indicators have been used to improve the accuracy of land cover change detection. However, these measurements, in terms of the single shape features, can hardly capture the complete trends of curves affected by the unsynchronized phenology. Therefore, the calculated change magnitudes are indistinct such that changes and no-changes have a low contrast. This problem has prevented traditional change detection methods from achieving a higher accuracy using bi-temporal images or NDVI time series. In this paper, a multiple shape parameters-based change detection method is proposed by combining the spectral correlation operator and the shape features of NDVI temporal curves (phase angle cumulant, baseline cumulant, relative cumulation rate, and zero-crossing rate). The change magnitude is derived by integrating all the inter-annual differences of these shape parameters. The change regions are discriminated by an automated threshold selection method known as histogram concavity analysis. The results showed that the mean differences in the change magnitudes of the proposed method between 2100 changed and 2523 unchanged pixels was 32%, the overall accuracy was approximately 88%, and the kappa coefficient was 0.76. A comparative analysis was conducted with bi-temporal image-based methods and NDVI time series-based methods, and we demonstrate that the proposed method is more effective and robust than traditional methods in achieving high-contrast change magnitudes and accuracy.

Forests play a vital role in combating gradual developmental deficiencies and balancing regional ecosystems, yet they are constantly disturbed by man-made or natural events. Therefore, developing a timely and accurate forest disturbance detection strategy is urgently needed. The accuracy of traditional detection algorithms depends on the selection of thresholds or the formulation of complete rules, which inevitably reduces the accuracy and automation level of detection. In this paper, we propose a new multitemporal convolutional network framework (MT-CNN). It is an integrated method that can realize long-term, large-scale forest interference detection and distinguish the types (forest fire and harvest/deforestation) of disturbances without human intervention. Firstly, it uses the sliding window technique to calculate an adaptive threshold to identify potential interference points, and then a multitemporal CNN network is designed to render the disturbance types with various disturbance duration periods. To illustrate the detection accuracy of MT-CNN, we conducted experiments in a large-scale forest area (about 990 ${\mathrm{km}}^2$) on the west coast of the United States (including northwest California and west Oregon) with long time-series Landsat data from 1986 to 2020. Based on the manually annotated labels, the evaluation results show that the overall accuracies of disturbance point detection and disturbance type recognition reach $90\%$. Also, this method is able to detect multiple disturbances that continuously occurred in the same pixel. Moreover, we found that forest disturbances that caused forest fire repeatedly appear without a significant coupling effect with annual temporal and precipitation variations. Potentially, our method is able to provide large-scale forest disturbance mapping with detailed disturbance information to support forest inventory management and sustainable development.