Nowadays, 3-D convolutional neural networks (3-D CNN) have attracted lots of attention in the spectral-spatial classification of hyperspectral imageries (HSI). In this model, the feed-forward processing structure reduces the computational burden of 3-D structural processing. However, this model as a vector-based methodology cannot analyze the full content of the HSI information, and as a result, its features are not quite discriminative. On the other hand, convolutional long short-term memory (CLSTM) can recurrently analyze the 3-D structural data to extract more discriminative and abstract features. However, the computational burden of this model as a sequence-based methodology is extremely high. In the meanwhile, the robust spectral-spatial feature extraction with a reasonable computational burden is of great interest in HSI classification. For this purpose, a two-stage method based on the integration of CNN and CLSTM is proposed. In the first stage, 3-D CNN is applied to extract low-dimensional shallow spectral-spatial features from HSI, where information on the spatial features are less than that of the spectral information; consequently, in the second stage, the CLSTM, for the first time, is applied to recurrently analyze the spatial information while considering the spectral one. The experimental results obtained from three widely used HSI datasets indicate that the application of the recurrent analysis for spatial feature extractions makes the proposed model robust against different spatial sizes of the extracted patches. Moreover, applying the 3-D CNN prior to the CLSTM efficiently reduces the model’s computational burden. The experimental results also indicated that the proposed model led to a 1% to 2% improvement compared to its counterpart models.

High-resolution urban image clustering has remained a challenging task. This is mainly because its performance strongly depends on the discrimination power of features. Recently, several studies focused on unsupervised learning methods by autoencoders to learn and extract more efficient features for clustering purposes. This paper proposes a Boosted Convolutional AutoEncoder (BCAE) method based on feature learning for efficient urban image clustering. The proposed method was applied to multi-sensor remote-sensing images through a multistep workflow. The optical data were first preprocessed by applying a Minimum Noise Fraction (MNF) transformation. Then, these MNF features, in addition to the normalized Digital Surface Model (nDSM) and vegetation indexes such as Normalized Difference Vegetation Index (NDVI) and Excess Green (ExG(2)), were used as the inputs of the BCAE model. Next, our proposed convolutional autoencoder was trained to automatically encode upgraded features and boost the hand-crafted features for producing more clustering-friendly ones. Then, we employed the Mini Batch K-Means algorithm to cluster deep features. Finally, the comparative feature sets were manually designed in three modes to prove the efficiency of the proposed method in extracting compelling features. Experiments on three datasets show the efficiency of BCAE for feature learning. According to the experimental results, by applying the proposed method, the ultimate features become more suitable for clustering, and spatial correlation among the pixels in the feature learning process is also considered.

Rational function model (RFM) is one of the most popular methods of geometrically correcting high-resolution satellite images (HRSIs). This model encounters overparameterization problem due to the existence of highly correlated RFM coefficients, namely, rational polynomial coefficients (RPCs). Recently, a number of methods have been proposed based on particle swarm optimization (PSO) to find the optimal RPCs. Although these algorithms are useful for determining the optimal RPCs, their results are strongly influenced by changes in both initial values and ground control points (GCPs) distribution. To address this problem, this study proposes a modified version of PSO based on the $k$ -fold cross-validation, known as PSO-KFCV, which works well even in the presence of limited GCPs. To evaluate the performance of the proposed method, four different HRSIs were used. Our experimental results indicate that PSO-KFCV is indeed robust against the initial values and GCPs distribution. In addition, the experiments demonstrated that the proposed method led to significant improvement with respect to state-of-the-art meta-heuristic methods.

Unsupervised or clustering algorithms can be considered to overcome the need for both high-quantity and high-quality training data for hyperspectral data classification. One of the most widely used algorithms for the clustering of remotely-sensed data is partitional clustering. Partitional clustering is affected by 1) the optimal number of clusters (NOC), 2) the position of cluster centers in hyper-dimension space, and 3) a set of optimally discriminating spectral bands. Among these three parameters, the NOC and their positions can be found simultaneously by dynamic clustering approaches. In this paper, an innovative two-stage dynamic clustering method is proposed and evaluated. In the first stage, the optimum set of solutions is achieved by a multi-objective particle swarm optimization. Then, using an efficient multi-criteria decision-making method, namely, the technique for order of preference by similarity to ideal solution (TOPSIS), a ranking is done among the optimal set of solutions to select the best one. Comparisons with some classic algorithms reveal that the proposed method is more effective at detecting the optimal number and position of clusters. In addition, the proposed algorithm generates better clustering results for hyperspectral data. Indeed, our method leads to a 5%-10% improvement upon classification accuracy.