While deep learning models have been extensively applied to land-use land-cover (LULC) problems, it is still a relatively new and emerging topic for separating and classifying wetland types. On the other hand, ensemble learning has demonstrated promising results in improving and boosting classification accuracy. Accordingly, this study aims to develop a classification system for mapping complex wetland areas by incorporating deep ensemble learning and satellite datasets. To this end, time series of Sentinel-1 dual-polarized Synthetic Aperture Radar (SAR) dataset, alongside Sentinel-2 multispectral imagery (MSI), are used as input data to the model. In order to increase the diversity of the extracted features, the proposed model, herein called WetNet, consists of three different submodels, comprising several recurrent and convolutional layers. Furthermore, multiple ensembling sections are added to different stages of the model to increase the transferability of the model (to other areas) and the reliability of the final results. WetNet is evaluated in a complex wetland area located in Newfoundland, Canada. Experimental results indicate that WetNet outperforms the state-of-the-art deep models (e.g., InceptionResnetV2, InceptionV3, and DenseNet121) in terms of both the classification accuracy and processing time. This makes WetNet an efficient model for large-scale wetland mapping application. The python code of the proposed WetNet model is available at the following link for the sake of reproducibility: https://colab.research.google.com/drive/1pvMOd3_tFYaMYGyHNfxqDxOiwF78lKgN?usp=sharing

An increasing availability of remote sensing data in the era of geo big-data makes producing well-represented, reliable training data to be more challenging and requires an excessive amount of human labor. In addition, the rapid increase in graphics processing unit processing power has enabled the development of advanced deep learning algorithms, which achieve impressive results in the field of satellite image processing. However, they require a huge and comprehensive training dataset to avoid overfitting problems and to represent a generalizable model. Thus, moving toward the development of nonsupervised deep learning (NSDL) models in different remote sensing applications is an inevitable need. To provide an initial response to that need, this article performs a comprehensive review and systematic meta-analysis of recently published research articles focusing on the applications of NSDL for remote sensing data processing. In order to identify future research directions and formulate recommendations, we extract trends and highlight interesting approaches from this large body of literature. Consequently, current challenges, prospects, and recommendations are also discussed to uncover the trend. According to the results, there is a sharp increasing trend in the applicability of NSDL methods during these few years particularly, with the advent of new deep architectures, such as adversarial, graph, and transformer models. As a result, this review article discusses different remote sensing data processing applications and challenges that can be addressed using NSDL approaches.

InSAR processing is vastly used for land deformation monitoring from the space. Machine learning methods are known as strong tools for data modeling as well as predicting. In this study, we are going to model and predict the future behavior of land subsidence by InSAR processing and leveraging deep learning methods over the lands in the vicinity of Lake Urmia (located in the northwest of Iran). Accordingly, Sentinel-1 data over 56 months from November 2014 to June 2019 and small baseline subsets (SBAS) InSAR methods were utilized. Several regions with a high rate of subsidence were identified (maximum monthly subsidence of 13.3 mm). Furthermore, environmental factors affecting subsidence were considered. Therefore, parameters such as rainfall, groundwater, and lake area variations were measured using TRMM, GRACE, and MODIS satellite data, respectively. In order to determine and assess the relation between land deformations and environmental variations, several machine learning methods were implemented. The environmental parameters were used as the input of models and ground deformations as the target to be predicted. Eventually, ground deformations were estimated using multi-layer perceptron (MLP), convolutional neural network (CNN), and long short-term memory (LSTM) networks, in which each network had strengths and weaknesses on different occasions. Thus, by blending the forecast of the three models, a weighted ensemble was constructed, which outperformed the single models and reached the root mean square error (RMSE), mean absolute error (MAE), and standard deviation (SD) of 8.2 mm, 6.4 mm, and ±5.2 mm, respectively. The result indicated that although each single model had proper accuracy, an ensemble model can improve land deformation anticipation using the strength of networks in various conditions.

Traditional mapping and monitoring of agricultural fields are expensive, laborious, and may contain human errors. Technological advances in platforms and sensors, followed by artificial intelligence (AI) and deep learning (DL) breakthroughs in intelligent data processing, led to improving the remote sensing applications for precision agriculture (PA). Therefore, technological advances in platforms and sensors and intelligent data processing methods, such as machine learning and DL, and geospatial and remote sensing technologies, have improved the quality of agricultural land monitoring for PA needs. However, providing ground truth data for model training is a time-consuming and tedious task and may contain multiple human errors. This paper proposes an automated and fully unsupervised framework based on image processing and DL methods for plant detection in agricultural lands from very high-resolution drone remote sensing imagery. The proposed framework’s main idea is to automatically generate an unlimited amount of simulated training data from the input image. This capability is advantageous for DL methods and can solve their biggest drawback, i.e., requiring a considerable amount of training data. This framework’s core is based on the faster regional convolutional neural network (R-CNN) with the backbone of ResNet-101 for object detection. The proposed framework’s efficiency was evaluated by two different image sets from two cornfields, acquired by an RGB camera mounted on a drone. The results show that the proposed method leads to an average counting accuracy of 90.9%. Furthermore, based on the average Hausdorff distance (AHD), an average object detection localization error of 11 pixels was obtained. Additionally, by evaluating the object detection metrics, the resulting mean precision, recall, and F1 for plant detection were 0.868, 0.849, and 0.855, respectively, which seem to be promising for an unsupervised plant detection method.

Hyperspectral imagery (HSI) creates a lot of applications in target or anomaly detection due to their rich spectral content. Generally, one scene of an HSI contains more than one class. Therefore, the Gaussian distribution assumption of the background fails. Furthermore, the high dimensionality of data makes background modelling more difficult by increasing redundancy and disturbances. In this paper, a segmented-distance based anomaly detection method is proposed for HSI. The proposed method is based on segmentation and takes advantage of the statistical properties of the segmented areas to suppress the false-alarms. In addition to that, nonlinear feature extraction based on convolutional stacked auto-encoder (SAE) neural networks are implemented to extract deep and nonlinear relations from the input data. Both 1-D and 2-D convolutional layers are investigated. The proposed method is tested on the three different datasets. The experimental results show that the integration of segmentation and deep feature extraction generally performs better than other state-of-the-art methods.