Description: Presents the front cover for this issue of the publication.

Description: We propose a novel spatiotemporal fusion method based on deep convolutional neural networks (CNNs) under the application background of massive remote sensing data. In the training stage, we build two five-layer CNNs to deal with the problems of complicated correspondence and large spatial resolution gaps between MODIS and Landsat images. Specifically, we first learn a nonlinear mapping CNN between MODIS and low-spatial-resolution (LSR) Landsat images and then learn a super-resolution CNN between LSR Landsat and original Landsat images. In the prediction stage, instead of directly taking the outputs of CNNs as the fusion result, we design a fusion model consisting of high-pass modulation and a weighting strategy to make full use of the information in prior images. Specifically, we first map the input MODIS images to transitional images via the learned nonlinear mapping CNN and further improve the transitional images to LSR Landsat images via the fusion model; then, via the learned SR CNN, the LSR Landsat images are supersolved to transitional images, which are further improved to Landsat images via the fusion model. Compared with the previous learning-based fusion methods, mainly referring to the sparse-representation-based methods, our CNNs-based spatiotemporal method has the following advantages: 1) automatically extracting effective image features; 2) learning an end-to-end mapping between MODIS and LSR Landsat images; and 3) generating more favorable fusion results. To examine the performance of the proposed fusion method, we conduct experiments on two representative Landsat–MODIS datasets by comparing with the sparse-representation-based spatiotemporal fusion model. The quantitative evaluations on all possible prediction dates and the comparison of fusion results on one key date in both visual effect and quantitative evaluations demonstrate that the proposed method can generate more accurate fusion results.

Description: Accurately monitoring forest dynamics in the tropical regions is essential for ecological studies and forest management. In this study, images from phase-array L-band synthetic aperture radar (PALSAR), PALSAR-2, and Landsat in 2006–2010 and 2015 were combined to identify tropical forest dynamics on Hainan Island, China. Annual forest maps were first mapped from PALSAR and PALSAR-2 images using structural metrics. Those pixels with a high biomass of sugarcane or banana, which are widely distributed in the tropics and subtropics and have similar structural metrics as forests, were excluded from the SAR-based forest maps by using phenological metrics from time series Landsat imagery. The optical–SAR-based forest maps in 2010 and 2015 had high overall accuracies (OA) of 92–97% when validated with ground reference data. The resultant forest map in 2010 shows good spatial agreement with public optical-based forest maps (OA = 88–90%), and the annual forest maps (2007–2010) were spatiotemporally consistent and more accurate than the PALSAR-based forest map from the Japan Aerospace Exploration Agency (OA = 82% in 2010). The areas of forest gain, loss, and net change on Hainan Island from 2007 to 2015 were 415 000 ha (+2.17% yr –1 ), 179 000 ha (–0.94% yr –1 ), and 236 000 ha (+1.23% yr –1 ), respectively. About 95% of forest gain and loss occurred in those areas with an elevation less than 400 m, where deciduous rubber, eucalyptus plantations, and urbanization expanded rapidly. This study demonstrates the potential of- PALSAR/PALSAR-2/Landsat image fusion for monitoring annual forest dynamics in the tropical regions.

Description: Tobacco plant detection plays an important role in the management of tobacco planting. In this paper, a new algorithm based on deep neural networks is proposed to detect tobacco plants in images captured by unmanned aerial vehicles (UAVs) (called UAV images). These UAV images are characterized by a very high spatial resolution (35 $text{mm}$ ), and consequently contain an extremely high level of detail for the development of automatic detection algorithms. The proposed algorithm consists of three stages. In the first stage, a number of candidate tobacco plant regions are extracted from UAV images with the morphological operations and watershed segmentation. Each candidate region contains a tobacco plant or a nontobacco plant. In the second stage, a deep convolutional neural network is built and trained with the purpose of classifying the candidate regions as tobacco plant regions or nontobacco plant regions. In the third stage, postprocessing is performed to further remove the nontobacco plant regions. The proposed algorithm is evaluated on a UAV image dataset. The experimental results show that the proposed algorithm performs well on the detection of tobacco plants in UAV images.

Description: Quantification of tree canopy area and aboveground biomass is essential for monitoring ecosystems’ ecological functionalities, e.g., carbon sequestration and habitat provision. Miombo woodlands are vastly existent in developing countries that often lack resources to acquire LiDAR data or high spatiospectral resolution remote sensing data that have been proven to accurately estimate these structural attributes. This study explored the utility of freely available (via Google Maps) high (1-m) resolution red, green, and blue (RGB) satellite imagery in combination with object-based image analysis (OBIA) for estimating tree canopy area and aboveground biomass in Miombo woodlands. We randomly established 41 225-m 2 plots in Mukuvisi Woodland, Zimbabwe, and used RGB data with OBIA to estimate tree canopy area in those plots. We also field measured the height, canopy area, and trunk diameter at breast height of all trees that fell in those plots, then used the field data and a published allometric equation to estimate aboveground tree biomass (AGB). OBIA classification accuracy was high (Jaccard similarity index = 0.96). Data analysis showed significant positive linear relationship between AGB and field-measured canopy area $(R^{2} = {{0.87}}, p 〈 {{0.003}})$ , and significant exponential relationships between: 1) OBIA-derived canopy area and AGB $(R^{2} = {{0.56}}, p 〈 {{0.0001}})$ ; and 2) field-measured canopy area and OBIA-derived canopy area $(R^{2} = {{0.63}}, p 〈 {{0.0001}})$ , and no significant differences $(t = {{19.67}}, df = {{78}}, p = {{0.28}})$ between field-measured canopy are- ( $bar{ times } = 187.11,{{rm{m}}^2},sigma = 127.03$ ) and OBIA-derived canopy area ( $bar{ times } = 163.00,{{rm{m}}^2},sigma = 50.08$ ). We conclude that RGB data with OBIA are suitable for estimating tree canopy area in Miombo woodlands for various low-accuracy purposes (e.g., biomass estimation).

Description: The fuzzy c-means (FCM) algorithm and many improved algorithms incorporating spatial information have been proven to be effective in image segmentation. However, these methods are not adaptable to process synthetic aperture radar (SAR) images owing to the intrinsic speckle noise. Our solution, which enables the effective segmentation of SAR images by guaranteeing noise-immunity and edge detail preservation simultaneously, is to propose a robust FCM algorithm based on Bayesian nonlocal spatial information (RFCM $_$ BNL). The nonlocal idea considers more useful information for generating an auxiliary image. We measure the similarity between patches by utilizing a dedicated noise model for SAR images, and then apply it to the Bayesian formulation. Then we derive a new statistical distance, which is insensitive to speckle noise. Additionally, we ensure that the algorithm is robust to outliers by employing the entropy of the local gray-level histogram to control the extent to which the nonlocal spatial information term is adaptive to pixels. Experiments on simulated and real SAR images show that RFCM $_$ BNL obtains the best result for SAR image segmentation compared with seven other fuzzy clustering algorithms.

Description: Continuous monitoring of topographic heights and changes in tidal flats is challenging, as it is generally difficult to observe topographic changes from on-site measurements or remote sensing techniques with high resolution and high accuracy. In this regard, an interferometric synthetic aperture radar (In-SAR) can be an effective tool to generate precise digital elevation models (DEMs) and detect large-scale topographic changes. Nevertheless, utilizing the In-SAR to detect topographic changes in tidal flats is not practical because the average slope of tidal flats is usually less than 5°, and the overall spatial and temporal variations of height are not significant. Therefore, the accuracy of In-SAR DEMs must be high to detect meaningful topographic changes. In order to minimize the error of In-SAR DEMs, height of ambiguity and random phase deviation of interferograms should be taken into account. These two factors are related to incidence angle and baseline. We simulated topographic error levels in tidal flats for a single-pass In-SAR system such as TanDEM-X. Phase error of interferograms was derived using the relationship between In-SAR coherence and the probability density function of phase deviation. Signal-to-noise ratio and geometric decorrelation were formulated by the function of baseline, incidence angle, and surface slope. The simulation results show that the height error of the DEM was minimized to lower than 15 cm when the baseline was 1500 m with an incidence angle of 29° in the TanDEM-X system. Finally, the validation of simulation results was carried out by comparing them with TanDEM-X DEM accuracies in tidal flats.

Description: Multilooking is a key step in interferometric processing, especially in so-called coherent stacking interferometry approaches. In the past, multilooking algorithms were mainly implemented in the spatial domain on single interferometric pairs. With continuous development in repeat pass capabilities, multitemporal coherent synthetic aperture radar (SAR) images are now generally acquired more easily, thus providing the possibility to exploit also the temporal signature for multilooking. In this field, the possibility to carry out adaptive multilooking is fundamental for the improvement of interferometric processing. A basic similarity test has been introduced in the SqueeSAR approach, namely the Kolmogorov–Smirnov test. Furthermore, similarity tests have been discussed in terms of real-valued data vector, so only amplitude information can be utilized: The influence of the phase signal is typically ignored. To fully exploit the complex information acquired by coherent SAR systems, this paper proposes an adaptive multilooking algorithm based on complex patch (AMCP). The complex signal, which is fundamental in interferometric systems, is here exploited for the derivation of a new patch selection method. The AMCP algorithm can be applicable to all multitemporal techniques that need filtering, including the InSAR stacks of single main image and multi main images. Experiments on simulated data and real data validate that the proposed algorithm has the highlighting major advantages in improving measurement precision compared with traditional methods.

Description: Tea is an important cash crop in Kenya, grown in a climatically restricted geographic area where climatic variability is starting to affect yield productivity levels. This paper assesses the feasibility of monitoring tea growth between, but also within fields, using X-band COSMO-SkyMed SAR images (five images at VV polarization and five images at HH polarization). We detect the harvested and nonharvested areas for each field, based on the loss of interferometric coherence between two images, with an accuracy of 52% at VV polarization and 74% at HH polarization. We then implement a normalization method to isolate the scattering component related to shoot growth and eliminate the effects of moisture and local incidence angle. After normalization, we analyze the difference in backscatter between harvested and nonharvested areas. At HH polarization, our backscatter normalization reveals a small decrease ( $sim0.1$  dB) in HH backscatter after harvest. However, this decrease is too small for monitoring shoot growth. The decrease is not clear at VV polarization. This is attributed to the predominantly horizontal orientation of the harvested leaves.

Description: This paper addresses the issue of deceptive jamming against synthetic aperture radar (SAR) by using 1-bit sampling and time-varying threshold (TVT). With 1-bit intercepted SAR signal, the multipliers involved in a convolution is replaced by xnor gates, which considerably simplify the jamming signal generation. Moreover, the TVT is used for 1-bit quantization before retransmission to retain the relative amplitude information of the jamming signal. As a result, the proposed deceptive jamming schemes are superior to their conventional counterpart in terms of realization. Effects of harmonics and oversampling are analyzed to evaluate the performance degradations caused by the 1-bit sampling and TVT. Simulation results are provided to confirm the validity of the proposed schemes.