Multi-class imbalanced problems have attracted growing attention from the real-world classification tasks in engineering. The underlying skewed distribution of multiple classes poses difficulties for learning algorithms, which becomes more challenging when considering overlapping between classes, lack of representative data and mixed-type data. In this work, we address this problem in a data-oriented way. Motivated by a recently proposed over-sampling technique designed for numeric data sets, Mahalanobis Distance-based Over-sampling (MDO), we use this technique to capture the covariance structure of the minority class and to generate synthetic samples along the probability contours for learning algorithms. Based on MDO, we further improve the over-sampling strategy and generalize it for mixed-type data sets. The established technique, Adaptive Mahalanobis Distance-based Over-sampling (AMDO), introduces GSVD (Generalized Singular Value Decomposition) for mixed-type data, develops a partially balanced resampling scheme and optimizes the sample synthesis. Theoretical analysis is conducted to demonstrate the reasonability of AMDO. Extensive experimental testing is performed on 15 multi-class imbalanced benchmarks and two data sets for precipitation phase recognition in comparison with several state-of-the-art multi-class imbalanced learning methods. The results validate the effectiveness and robustness of our proposal.

In recent years, many models based on the convolutional neural network have achieved high-quality reconstruction for single image super-resolution. Meanwhile, many researches on image super-resolution have been applied to various fields. However, only a few research works have been applied to climate prediction. In this paper, we present ResLap to achieve high-resolution climate prediction. ResLap is a spatial downscaling method that converts low spatial resolution climate data into high-resolution regional climate forecasts. This method mainly introduces a novel residual dense block (RDB) into the Laplacian pyramid super-resolution network (LapSRN). Among them, we use LapSRN to achieve upsampling image reconstruction, and adopt RDB to fully extract the hierarchical features from all the convolutional layers. Extensive experimental results on benchmark climate datasets show that our new proposed model performs better than many super-resolution methods. Besides, the climate data are more complicated than the general image, because of its dynamic and chaotic nature. To facilitate model training, we integrate original climate data provided by the China Meteorological Administration, then convert it into trainable climate images. We also publish some climate image datasets online for research. Finally, we avoid the checkerboard artifacts in the generated high-resolution climate images.

Climate prediction is susceptible to a variety of meteorological factors, and downscaling technology is used for high-resolution climate prediction. This technology can generate small-scale regional climate prediction from large-scale climate output information. Inspired by the concept of image super resolution, we propose to apply the convolutional neural network (CNN) to downscaling technology. However, some unpleasant artifacts always appear in the final climate images generated by existing CNN-based models. To further eliminate these unpleasant artifacts, we present a new training strategy for the generative adversarial network, termed DeepDT. The key idea of our DeepDT is to train a generator and a discriminator separately. More specifically, we apply the residual-in-residual dense block as the basic frame structure to fully extract the features of the input. Additionally, we innovatively use a CNN model to fuse multiple climate elements to generate trainable climate images, and build a high-quality climate data set. Finally, we evaluate the DeepDT using the proposed climate data sets, and the experiments indicate that DeepDT performs best compared to most CNN-based models in climate prediction.

Weather forecasting, which is challenging due to the complex atmospheric correlation, focuses on providing explicit meteorological estimations as accurate as possible. Recently, techniques based upon convolutional-recurrent networks have shown dramatic performance in domains including radar echo prediction and precipitation forecasting, indicating that deep learning models have great potential in this area. However, existing methods concentrate on adding extra paralleled memory cells to the inner recurrent unit, where the information is mutually independent. To extract spatial–temporal features with stronger correlation, this letter introduced an axial attention memory module with quasi state-in-state cascaded manner. Benefitting from this unit, the spatial–temporal features can be aggregated and embedded into standard ConvGRU formulating Axial aTtention Memory Cascaded ConvGRU (ATMConvGRU) for sequential weather prediction. Experimental results compared with many state-of-the-art methods on four types of weather forecasting related datasets, including temperature, relative humidity, wind, and radar echo, demonstrate the effectiveness and superiority of the proposed ATMConvGRU.

Precipitation downscaling, which is similar to the mechanism of single-image super-resolution (SR), aims to improve the spatial resolution of rain maps. It is of great practical value and theoretical significance. This letter presents a new deep precipitation downscaling (DPD) method, named auxiliary guided spatial distortion (AGSD) network, motivated by SR techniques. Specifically, an auxiliary guided module (AGM), which takes multiple meteorological elements (e.g., temperature, relative humidity, and wind) as input, is proposed for getting more accurate rain map features. Meanwhile, a simple but effective spatial distortion module (SDM) is proposed. Benefitting from SDM, the DPD model can rectify the rain map via terrain correlation. Furthermore, to improve the model performance among various rain intensity (including small rain, moderate rain, heavy rain, and storm), a threat score-driven pseudo threat-score (PTS) loss is presented. Experimental results compared with state-of-the-art methods demonstrate the superiority of the proposed method.

Image-based quantified visibility estimation is an important task for both atmospheric science and computer vision. Traditional methods rely largely on meteorological observation or manual camera calibration, which restricts its performance and generality. In this paper, we propose a new end-to-end pipeline for single image-based quantified visibility estimation by an elaborate integration between meteorological physical constraint and deep learning architecture design. Specifically, the proposed Deep Quantified Visibility Estimation Network (abbreviated as DQVENet) consists of three modules, i.e., the Transmission Estimation Module (TEM), the Depth Estimation Module (DEM), and the Extinction coEfficient Estimation Module (E3M). Casting on these modules, the meteorological prior constraint can be combined with deep learning. To validate the performance of DQVENet, this paper also constructs a traffic image dataset (named QVEData) with accurate visibility calibration. Experimental results compared with many state-of-the-art methods on QVEData demonstrate the effectiveness and superiority of DQVENet.